1,2-Hydroxypyridonates as contrast agents for magnetic resonance imaging: TREN-1,2-HOPO

Article abstract of DOI:10.1021/ic700985j

1,2-Hydroxypyridinones (1,2-HOPO) form very stable lanthanide complexes that may be useful as contrast agents for magnetic resonance imaging (MRI). X-ray diffraction of single crystals established that the solid-state structures of the Eu(III) and the pre

Full text of DOI:10.1021/ic700985j

Inorg. Chem. 2007, 46, 9182−9191
1,2-Hydroxypyridonates as Contrast Agents for Magnetic Resonance
Imaging: TREN-1,2-HOPO
Christoph J. Jocher,† Evan G. Moore,† Jide Xu,† Stefano Avedano,‡ Mauro Botta,‡ Silvio Aime,§ and
Kenneth N. Raymond*,†
Department of Chemistry, UniVersity of California, Berkeley, California 94720-1460, Dipartimento
di Scienze dell’Ambiente e della Vita, UniVersita` del Piemonte Orientale “A. AVogadro”, Via
Bellini 25/G, I-15100 Alessandria, Italy, and Dipartimento di Chimica I. F. M., UniVersita` di
Torino, Via P. Giuria 7, I-10125 Torino, Italy
Received May 21, 2007
1,2-Hydroxypyridinones (1,2-HOPO) form very stable lanthanide complexes that may be useful as contrast agents
for magnetic resonance imaging (MRI). X-ray diffraction of single crystals established that the solid-state structures
of the Eu(III) and the previously reported [Inorg. Chem. 2004, 43, 5452] Gd(III) complex are identical. The recently
discovered sensitizing properties of 1,2-HOPO chelates for Eu(III) luminescence [J. Am. Chem. Soc. 2006, 128,
10 067] allow for direct measurement of the number of water molecules coordinated to the metal center. Fluorescence
measurements of the Eu(III) complex corroborate that, in solution, two water molecules coordinate the lanthanide
(q
anion binding interactions of lanthanide TREN-1,2-HOPO complexes in solution, studied by relaxivity, revealing
only very weak oxalate binding (KA 82.7
6.5 M-1). Solution thermodynamic studies of the metal complex and
) 2) as proposed from the analysis of NMRD profiles. In addition, fluorescence measurements have verified the
)
±
free ligand have been carried out using potentiometry, spectrophotometry, and fluorescence spectroscopy. The
metal ion selectivity of TREN-1,2-HOPO supports the feasibility of using 1,2-HOPO ligands for selective lanthanide
binding [pGd
) 19.3 (2), pZn ) 15.2 (2), pCa ) 8.8 (3)].
Introduction
(III) nitrogen bond does not significantly contribute to chelate
stability in these compounds. Thus, for aminocarboxylate
MRI-CA only octadentate ligands provide sufficient chelate
stability, which limits the number of coordinated water
molecules to one (q ) 1). The resulting limit in relaxivity
can be circumvented by utilizing the known oxophilicity of
lanthanides. Purely oxygen donor ligands such as hydroxy-
pyridinone (HOPO, Chart 1) and terephthalamide (TAM)
chelates allow two5 or three6 coordinated water molecules
(q ) 2 or 3) at the metal center without destroying chelate
stability and have been established as candidates for high
relaxivity next generation contrast agents. These ligands
combine enhanced relaxivity and high stabilitysa combina-
tion considered impossible for small molecular weight MRI-
CA based on aminocarboxylate ligands.7 The further devel-
High chelate stability is essential for lanthanide complexes
in medicine because lanthanide ions have known toxicity
via their interaction with Ca(II) binding sites.2 Additionally,
the application of weak Gd(III) chelators as MRI contrast
agent (MRI-CA) can cause nephrogenic systemic fibrosis in
patients with low kidney activity.3 To date, clinically used
gadolinium based contrast agents (MRI-CA) rely on conve-
niently available aminocarboxylate chelates derived from
DTPA and macrocyclic DOTA.4 However, the weak Gd-
* To whom correspondence should be addressed. Phone: 510-642-7219.
Fax: 510-486-5283. E-mail: raymond@socrates.berkeley.edu.
† University of California.
‡ Universita` del Piemonte Orientale “A. Avogadro”.
§ Universita` di Torino.
(1) High Relaxivity Gadolinium MRI Agents. 22. Part 21: Jocher, C. J.;
Botta, M.; Avedano, S.; Moore, E. G.; Xu, J.; Raymond, K. N. Inorg.
Chem. 2007, 46, 4796-4798.
(4) Caravan, P.; Ellison, J. J.; Murray, T. J.; Lauffer, R. B. Chem. ReV.
1999, 99, 2293-2352.
(2) Cacheris, W. P.; Quay, S. C.; Rocklage, S. M. Magn. Reson. Imaging
1990, 8, 467-481.
(5) Xu, J.; Franklin, S. J.; Whisenhunt, Jr., D. W.; Raymond, K. N. J.
Am. Chem. Soc. 1995, 117, 7245-7246.
(3) Marckmann, P.; Skov, L.; Rossen, K.; Dupont, A.; Damholt, M. B.;
Heaf, J. G.; Thomsen, H. S. J. Am. Soc. Nephrol. 2006, 17, 2359-
2362.
(6) Werner, E. J.; Avedano, S.; Botta, M.; Hay, B. P.; Moore, E. G.; Aime,
S.; Raymond, K. N. J. Am. Chem. Soc. 2007, 129, 1870-1871.
(7) Raymond, K. N.; Pierre, V. C. Bioconjugate Chem. 2005, 16, 3-8.
9182 Inorganic Chemistry, Vol. 46, No. 22, 2007
10.1021/ic700985j CCC: $37.00
© 2007 American Chemical Society
Published on Web 10/03/2007

[Gd(TREN-1,2-HOPO)(H2O)2] for MRI
Chart 1. 1,2-HOPO, 3,2-HOPO, and TAM Chelates Are Used To
Prepare Stable Gd(III) Complexes as MRI Contrast Agents
time, fluorescence spectroscopy of a TREN-HOPO chelate
as the Eu(III) complex are used to support solution structure
assignments based on relaxivity measurements for the same
ligand with Gd(III). This comparative study of relaxometric
data of the Gd(III)-TREN-1,2-HOPO complex and lumines-
cence data of the isostructural Eu(III) complex was also
applied for determining binding affinities of bioavailable and
synthetically interesting anions for the Eu(III) and Gd(III)
complexes of TREN-1,2-HOPO and to assign the structure
of ternary complexes in solution. In addition, solution
thermodynamic stability was assessed for an estimate of
potential toxicity of TREN-1,2-HOPO chelates.
opment of mixed 3,2-HOPO-TAM chelates follows the
current trend of improving relaxometric properties of MRI-
CA by formation of polymers,8 dendrimers,9 or supramo-
lecular compounds.10 The feasibility of applying such 3,2-
HOPO chelates as MRI-CA has been proven successfully
in vivo.11
Experimental Details
General. All chemicals were used as received from Aldrich
unless otherwise noted. TREN-1,2-HOPO15 and 3-benzyloxy-4-
carboxy-2-pyridone (3,2-HOPO(Bn) acid)5 were synthesized as
1
previously reported. H and 13C NMR spectra were obtained on
One of the few disfavored features of 3,2-HOPO has been
that the solution structure of its lanthanide complexes has
not been elucidated by means other than relaxometry.
Methods such as Dy(III) induced paramagnetic shift in 17O
NMR spectroscopy12 or fluorescence spectroscopy of parent
Eu(III) or Tb(III) chelates13 have been investigated without
much success. As such, information about the solution
structure including the hydration number q for 3,2-HOPO
and TAM chelates has been inferred only by relaxometry
based on reasonable assumptions about the values of several
parameters and, whenever possible, on the comparison of
1H and 17O NMR relaxation data. The recent discovery of
the excellent luminescence sensitizing properties of 6-car-
boxylamide-1,2-hydroxypyridonates (1,2-HOPO) for Eu-
(III)14 allows the correlation of relaxometric data with
fluorescence spectroscopy for 1,2-HOPO compounds. Con-
trary to several well studied 3,2-HOPO chelates, only TREN-
1,2-HOPO15 and TACN-1,2-HOPO6 have been described as
1,2-HOPO chelates for MRI contrast agents. Comparison of
the solid-state structure and NMRD profile of Gd-TREN-
1,2-HOPO with Gd-TREN-3,2-HOPO led to the conclusion
that these complexes behave very similarly in solution.15
Thus, 1,2-HOPO can be used as a fluorescent tool to
elucidate solution structure of HOPO chelates by fluores-
cence spectroscopy in addition to relaxometry and solution
thermodynamics. The following contribution presents details
on the solution stability and structure of the Eu(III), Gd-
(III), and Zn(II) complexes of TREN-1,2-HOPO. For the first
Bruker AVQ-400 spectrometers and are reported in ppm. Mass
spectra and elemental (CHN) analyses were performed at the
Elemental Analysis Facility, College of Chemistry, UC Berkeley.
Unless specified, solvents were removed by rotary evaporation.
Synthesis. 3-Benzyloxy-4-methylcarbamido-1-methyl-2-pyri-
done. 3-Benzyloxy-4-carboxy-1-methyl-2-pyridone (2.56 g, 10
mmol) was dissolved in toluene (10 mL). Oxalylchloride (3.81 g,
30 mmol) and three drops of DMF were added. Volatiles were
removed after gas evolution ceased (∼1 h), and the remaining solid
was dissolved in dry THF (50 mL). Methylamine (40% in water;
50 mL) was diluted with THF (50 mL) and cooled on in ice bath,
and the above solution of 3-benzyloxy-4-carboxychloro-1-methyl-
2-pyridone was added. The reaction mixture was stirred at 0 °C
for 4 h, and dichloromethane (250 mL) was added. The reaction
mixture was washed with 1 M HCl (3 × 100 mL), 1 M KOH (3 ×
100 mL), and water (1 × 50 mL) and dried over Na2SO4. Volatiles
were removed under reduced pressure, and a yellowish solid
1
remained. Yield: 2.34 g (8.59 mmol, 86%). H NMR (CDCl3,
400.132 MHz) δ ) 7.89 (s, 1 H, amide NH), 7.43-7.33 (m, 5 H,
3
Bn-CH), 7.13 (d, 1 H, 6-CHHOPO, JHH ) 7.2 Hz), 6.78 (d, 1 H,
3
6-CHHOPO, JHH ) 7.2 Hz), 5.54 (s, 2 H, OCH2Ph), 3.59 (s, 3 H,
N
HOPOCH3), 2.78 (d, 3 H, 4-CONHCH3, 2JHH vic) 5.2 Hz). 13C NMR
(CDCl3, 100.623 MHz) δ ) 163.8 (CONH), 159.7 (CONH), 146.4,
136.2, 132.2, 130.8 (ipso-C-Bn), 129.1 (Ph-CH), 128.9 (Ph-CH),
128.8 (Ph-CH), 128.4, 126.9, 104.9 (5-CHHOPO), 65.1 (OCH2Ph),
37.7 (1-NCH3), 26.3 (4-CONHCH3). ESI-MS (MeOH; C15H16N2O3;
fw ) 272.30 g/mol) m/z ) 567.2 (95%, [2M + Na]+), 295.2 (100%,
[MNa]+).
3-Hydroxy-4-methylcarbamido-1-methyl-2-pyridone. 3-Ben-
zyloxy-4-methylcarbamido-1-methyl-2-pyridone (2.34 g, 8.59 mmol)
was dissolved in acetic acid (6 mL). Hydrochloric acid (37%; 6
mL) was added, and the solution was stirred for 18 h. Solvents
were removed under reduced pressure. Water (100 mL) was added
to the remaining yellowish suspension. The mixture was washed
with methylene chloride (2 × 100 mL; the organic phase contains
impurities and unreacted protected ligand). The remaining colorless
suspension was filtered to isolate the first crop of pure product.
Concentration of the aqueous filtrate allowed the collection of a
second crop of the title compound as colorless needles. Combined
yield: 658 mg (3.61 mmol, 42%). Melting point: 201 °C. fw )
182.18 g/mol. Elemental analysis for C8H10N2O3 calcd (found): C
(8) Allen, M. J.; Raines, R. T.; Kiessling, L. L. J. Am. Chem. Soc. 2006,
128, 6534-6535.
(9) Pierre, V. C.; Botta, M.; Raymond, K. N. J. Am. Chem. Soc. 2005,
127, 504-505.
(10) Pierre, V. C.; Botta, M.; Aime, S.; Raymond, K. N. J. Am. Chem.
Soc. 2006, 128, 9272-9273.
(11) Thompson, M. K.; Misselwitz, B.; Tso, L. S.; Doble, D. M. J.; Schmitt-
Willich, H.; Raymond, K. N. J. Med. Chem. 2005, 48, 3874-3877.
(12) Peters, J. A.; Huskens, J.; Raber, D. J. Prog. Nucl. Magn. Reson.
Spectrosc. 1996, 28, 283-350.
(13) De Horrocks, W. W., Jr.; Sudnick, D. R. J. Am. Chem. Soc. 1979,
101, 334-340.
(14) Moore, E. G.; Xu, J.; Jocher, C. J.; Werner, E. J.; Raymond, K. N. J.
Am. Chem. Soc. 2006, 128, 10067-10068.
1
52.74 (52.37), H 5.53 (5.59), 15.38 (14.99). H NMR (d6-DMSO,
(15) Xu, J.; Churchill, D. G.; Botta, M.; Raymond, K. N. Inorg. Chem.
2004, 43, 5492-5494.
400.132 MHz) δ ) 11.7 (s br, 1 H, OH), 8.40 (d, 1 H, amide NH),
Inorganic Chemistry, Vol. 46, No. 22, 2007 9183

Jocher et al.
3
7.15 (d, 1 H, 6-CHHOPO, JHH ) 7.2 Hz), 6.46 (d, 1 H, 6-CHHOPO
,
solutions containing constant ligand, metal, buffer [0.01 M of either
MES ()2-(N-morpholine)ethanesulfonic acid) pH 6.1; HEPES ()4-
(2-hydroxyethyl)-1-piperazineethanesulfonic acid) or TRIS ()
2-amino-2-hydroxymethylpropane-1,3-diol) for pH 7.4], and elec-
trolyte (0.1 M KCl) concentrations. If necessary, the pH of the
solutions was adjusted with concentrated HCl and/or KOH and the
solutions were diluted to identical volumes in buffered electrolyte
solution. The concentrations of TREN-1,2-HOPO relative to DTPA
ranged from 10:1 to 1:1000. After equilibrating for 24 h to ensure
thermodynamic equilibrium, concentrations of free and complexed
TREN-1,2-HOPO were determined from the absorption spectra,
using the wavelength range from 320 to 360 nm (pGd; Figures 5
and S7, Supporting Information), 320 to 350 nm (pZn; Figures 7
and S8, Supporting Information), and 320 to 350 nm (pCa; Figure
S9, Supporting Information) and the spectra of free and fully
complexed TREN-1,2-HOPO as references.
3JHH ) 7.2 Hz), 3.43 (s, 3 H, NHOPOCH3), 2.77 (d, 3 H,
4-CONHCH3, JHH vic) 5.2 Hz). 13C NMR (d6-DMSO, 100.623
2
MHz) δ ) 166.6 (CONH), 158.4 (CONH), 148.3, 128.1, 117.3,
102.7 (5-CHHOPO), 37.2 (1-NCH3), 26.5 (4-CONHCH3). FAB-MS
(NBA; fw ) 182.18 g/mol) m/z ) 183 (100%, [MH]+), 152 (43%,
[MH - CH3N]+), 136 (18%, [MH - CH3NO]+).
[Eu(TREN-1,2-HOPO)(H2O)2]. To a solution of TREN-1,2-
HOPO (61 mg, 0.10 mmol) in methanol (10 mL) was added a
solution of europium(III) chloride hexahydrate (36 mg, 0.1 mmol)
in methanol (10 mL) while being stirred. The clear solution became
turbid after 2 drops of dry pyridine were added. The mixture was
refluxed overnight under nitrogen, during which time the complex
precipitated as a white powder. The precipitate was collected, rinsed
with cold methanol, and dried to give the title complex (63 mg,
89%) as a white solid. Elemental analysis for EuC24H24N7O9‚H2O,
Calcd (Found): C, 39.79 (40.01); H, 3.62 (3.47); N, 13.53 (13.26).
FAB-MS (NBA; EuC24H24N7O9 fw ) 707.8) m/z ) 708 and 710
(MH+, 100%).
Photophysical Measurements. UV-visible absorption spectra
were recorded on a Varian Cary 300 double beam absorption
spectrometer using quartz cells of 1.00 cm path length. Emission
spectra were acquired on a HORIBA Jobin Yvon IBH FluoroLog-3
spectrofluorimeter, equipped with a 3 slit double grating excitation
and emission monochromators (2.1 nm/mm dispersion, 1200
grooves/mm). Spectra were reference corrected for both the
excitation light source variation (lamp and grating) and the emission
spectral response (detector and grating).
Crystals of [Eu(C24H24N7O9‚C3H7NO)]2 × 0.5C3H7NO ×
0.5C4H10O suitable for X-ray diffraction were obtained from vapor
diffusion of diethyl ether into a DMF solution (Table S1, Supporting
Information).
Solution Thermodynamics. The experimental protocols and
equipment used have been previously described.16 To determine
the protonation constants of the free ligand, approximately 13.5-
25.2 mg of ligand was dissolved in 25 or 50 mL of a 0.1 M aqueous
solution of KCl in a titration vessel (ligand concentration ∼0.4-
0.7 mM). Protonation constants of TREN-1,2-HOPO were deter-
mined by potentiometric (pH vs total proton concentration) titrations
using Hyperquad17 for data refinement. Each protonation constant
determination is the result of at least three experiments (where each
experiment consists of two titrations, the first one titrated with acid,
followed by a reverse titration with base). The equilibration time
between additions of titrant was 90 s for ligand titrations. Spec-
trophotometric titrations of ∼40 µmol solutions of TREN-1,2-
HOPO were fitted using pHAb18 for computation. Molar absor-
Luminescence lifetimes were determined on a HORIBA Jobin
Yvon IBH FluoroLog-3 spectrofluorimeter, adapted for time-
correlated single photon counting (TCSPC) and multichannel scaling
(MCS) measurements. A submicrosecond Xenon flashlamp (Jobin
Yvon, 5000XeF) was used as the light source, with an input pulse
energy (100 nF discharge capacitance) of ca. 50 mJ, yielding an
optical pulse duration of less than 300 ns at fwhm. Spectral selection
was achieved by passage through a double grating excitation
monochromator (2.1 nm/mm dispersion, 1200 grooves/mm). Emis-
sion was monitored perpendicular to the excitation pulse, again with
spectral selection achieved by passage through a double grating
excitation monochromator (2.1 nm/mm dispersion, 1200 grooves/
mm). A thermoelectrically cooled single photon detection module
(HORIBA Jobin Yvon IBH, TBX-04-D) incorporating fast rise time
PMT, wide bandwidth preamplifier, and picosecond constant
fraction discriminator was used as the detector. Signals were
acquired using an IBH DataStation Hub photon counting module,
and data analysis was performed using the commercially available
DAS 6 decay analysis software package from HORIBA Jobin Yvon
IBH. Goodness of fit was assessed by minimizing the reduced chi
squared function, ø2, and a visual inspection of the weighted
residuals. Each trace contained 10 000 points, and the reported
lifetime values result from three independent measurements. Typical
sample concentrations for both absorption and fluorescence mea-
surements were ca. 10-5-10-6 M and 1.0 cm cells in quartz suprasil
or equivalent were used for all measurements. An example of a
typical lifetime data trace and the corresponding best fit is shown
in Figure S1, Supporting Information.
bances of the species L3-, LH2-, LH2-, LH3, and LH4 were
+
determined (Figure S3, Supporting Information) and interpreted to
assign protonation sites. The smallest change in absorption between
two species results from the protonation at the amine nitrogen atom,
which is not directly adjacent to the 1,2-HOPO chromophores
monitored in the wavelength region between 280 and 380 nm.
1H (Table S6, Supporting Information; Figure 4) and 13C NMR
(Table S7 and Figure S3, Supporting Information) titrations were
performed on a solution of TREN-1,2-HOPO (88.34 mM) in 4 mL
of D2O. The pD of this solution was also monitored and varied by
stepwise addition of KOD/D2O (30%; ∼1-10 µL). The corre-
sponding pH was calculated by the relation pD ) 1.044 pH -
0.32.19 A small amount of methanol was used as an internal standard
(δ1H ) 3.32 ppm; δ13C ) 49.50 ppm).20
The general procedure used to determine the conditional stability
constants pGd, pZn or pCa of a TREN-1,2-HOPO complex was
competition batch titration adapted from a previous report.21 Varying
volumes of a standardized DTPA stock solution were added to
Relaxometric Measurements. The water proton 1/T1 longitu-
dinal relaxation rates (20 MHz, 25 °C) were measured on a Stelar
Spinmaster spectrometer (Mede, Pv, Italy) on 30-45 µM aqueous
solutions of the complexes. For the T1 determinations of solutions
the standard inversion-recovery method was used with a typical
90° pulse width of 3.5 µs, 16 experiments of 4 scans. The
reproducibility of the T1 data was estimated to be (1%. The
(16) Johnson, A. R.; O’Sullivan, B.; Raymond, K. N. Inorg. Chem. 2000,
39, 2652-2660.
(17) Gans, P.; Sabatini, A.; Vacca, A. Talanta 1996, 43, 1739.
(18) Gans, P.; Sabatini, A.; Vacca, A. Ann. Chim. (Rome) 1999, 89, 45-
49.
(19) Perin, D. D.; Dempsey, B. Buffers for pH and Metal Ions Control;
Chapman and Hall: London, 1974.
(20) Gottlieb, H.; Kotlyar, V.; Nudelman, A. J. Org. Chem. 1997, 62,
7512-7515.
(21) Doble, D. M. J.; Melchior, M.; O’Sullivan, B.; Siering, C.; Xu, J.;
Pierre, V. C.; Raymond, K. N. Inorg. Chem. 2003, 42, 4930-4937.
9184 Inorganic Chemistry, Vol. 46, No. 22, 2007

[Gd(TREN-1,2-HOPO)(H2O)2] for MRI
Figure 1. ORTEP plot (50% probability) of [Eu(TREN-1,2-HOPO)-
(DMF)]2. Noncoordinating solvent molecules and hydrogen atoms have been
omitted. Selected bond lengths (Å) and angles (deg): Eu1-O1a 2.389(4);
Eu1-O2a 2.387(4); Eu1-O1b 2.360(4); Eu1-O2b 2.366(4); Eu1-O1c
2.407(4); Eu1-O2c 2.379(4); Eu1-O1s1 2.396(4); Eu1-O3c* 2.467(4);
Eu1-O1b-N1b 119.2(3); O1b-N1b-C1b 115.9(5); Eu1-O2b-C1b
123.0(4); N1b-C1b-O2b 116.1(5); O1b-Eu1-O2b 64.73(14); Eu1-
O3c*-C6c* 125.8(4). (*) and (_3) indicate symmetry generated atoms.
Figure 2. The emission spectrum of a [Eu(TREN-1,2-HOPO)(H2O)2] in
aqueous solution at pH 6. The inset shows the expansion of emission spectra
in the 5D0 f 7FJ (J ) 0, 1; 570-600 nm) region and corresponding spectral
deconvolution into several overlapping Voigt functions.
Table 1. Solution Thermodynamic Constants of TREN-1,2-HOPO
logâ
pKa
7.16 (1)
5.62 (2)
4.32 (3)
3.60 (4)
pGd: 19.3 (2)
5.0 (2)
3.5 (5)
pZn: 15.2 (2)
pCa: 8.8 (3)
temperature was controlled with a Stelar VTC-91 air-flow heater
equipped with a copper-constantan thermocouple (uncertainty of
0.1(°C). In a typical titration experiment several (5-8) aqueous
solutions at pH ) 6.0 of the paramagnetic complex were prepared
containing different concentrations of the anionic species (0-0.04
M), and the water proton relaxation rate of each solution was
measured at 25 °C. The starting pH was adjusted by either HCl or
KOH. Moreover, the pH of the solutions was controlled before and
after the measurement.
L3- / HL2-
7.16 (1)
12.78 (2)
17.10 (3)
20.69 (3)
18.5 (2)
23.5 (2)
27.0 (5)
14.4 (2)
8.0 (3)
HL2- / H2L-
H2L- / H3L
H3L / H4L+
Gd + L / GdL
GdL + H / GdLH
GdLH + H / GdLH2
Zn + L / ZnL
Ca + L / CaL
∼ 6 in aqueous solution and in D2O, with lifetimes of 222
µs and 372 µs, respectively (Figure S1, Supporting Informa-
tion). Utilizing the improved Horrocks relationship given by
Parker24 gave an estimate of q ) 1.9, which confirms the
results obtained for the Gd(III) complex via relaxivity.
Results and Discussion
Solid-State Structure. The Eu(III) complex of TREN-
1,2-HOPO (Figure 1) is isostructural with the Gd(III)
complex. Both complexes crystallize in the monoclinic space
group P2/n in dinuclear units, linked via an amide carbonyl
oxygen atom and the lanthanide of the adjacent complex.
Each ligand coordinates one metal center in a hexadentate
fashion with a single DMF molecule completing the coor-
dination sphere of the metal to eight (CN ) 8). Distances
for Ln-O bonds vary only to a very small degree reflecting
the lanthanide contraction. Also the geometry around Gd-
(III) and Eu(III) atoms is identical as determined by the shape
measure S22 for eight-coordinate metal centers (Table S2,
Supporting Information) according to Kepert.23 The coordi-
nation geometry is described best by C2V symmetry corre-
sponding to a bicapped trigonal prism.
Solution Structure. In aqueous solution, the amide oxygen
atoms of DMF and of the adjacent ligand are replaced with
two fast exchanging water molecules (q ) 2) as suggested
by the NMRD profiles.15 For the first time, the number of
coordinated water molecules q of a TREN-HOPO chelate
could also be confirmed by luminescence spectroscopy of
its parent Eu(III) complex. The luminescence quenching of
coordinated water molecules can be used to determine q
directly.13 Luminescence decay times of the Eu-TREN-1,2-
HOPO complex were found to be monoexponential at pH
5
7
For Eu(III), the D0 f FJ transitions are amenable to
structural interpretation since, depending on the symmetry,
crystal field splitting yields a maximum of 2J + 1 sublevels.
Hence, emission spectra of the complex will reflect the
symmetry of its immediate surroundings,25 and a complete
analysis of crystal field splitting parameters for the Eu(III)
ion under various symmetry conditions using the principles
of group theory has been detailed in the literature (Table
S3, Supporting Information).26
Given the determination of an overall eight-coordinate (q
) 2) complex, a closer analysis of the emission spectra, in
5
7
particular for the D0 f FJ (J ) 0, 1) transitions, reveals
insight into the solution behavior of the Eu(III) complex
(Figure 2). The intensity of the J ) 0 band at ca. 17 300
cm-1 is comparable to J ) 1 bands, which suggests this
transition is symmetry allowed, implying C2V symmetry.
Further analysis for the J ) 1 region (ca. 16 700-17 100
cm-1) confirms this assignment since only under C2V sym-
(24) Beeby, A.; Clarckson, I. M.; Dickins, R. S.; Faulkner, S.; Parker, D.;
Royle, L.; de Sousa, A. S.; Williams, J. A. G.; Woods, M. J. Chem.
Soc., Perkin Trans. 1999, 2, 493-503.
(25) Lanthanide Probes in Life, Chemical and Earth Sciences: Theory and
Practice; Bu¨nzli, J. C. G., Choppin, G. R., Eds.; 1989.
(22) Xu, J.; Radkov, E.; Ziegler, M.; Raymond, K. N. Inorg. Chem. 2000,
39, 4156-4164.
(23) Kepert, D. L. Prog. Inorg. Chem. 1978, 24, 179.
(26) (a) Murray, G. M.; Sarrio, R. V.; Peterson, J. R. Inorg. Chim. Acta
1990, 176, 233-240. (b) Brecher, C.; Samelson, H.; Lempicki, A. J.
Chem. Phys. 1965, 42, 1081-1096.
Inorganic Chemistry, Vol. 46, No. 22, 2007 9185

Jocher et al.
Figure 3. Left: Spectrophotometric titration of TREN-1,2-HOPO. Right: Fitted molar absorbances for the five protonation states of TREN-1,2-HOPO.
Figure 4. 1H NMR titration of TREN-1,2-HOPO from pH (calculated from pD) 1.11-8.93 in D2O. The labeling scheme is depicted in Chart 2. * trace of
methanol for calibration.
metry are three J ) 1 bands expected, as clearly observed
in the spectral deconvolution. Hence, the site symmetry C2V
observed by X-ray crystallography for the TREN derivative
is maintained in solution.
(Figures S4-S6, Supporting Information) spectra gave
insight into molecular processes involved with protonation.
While only minor shifts are observed for the aromatic proton
resonances, the signals for 3′-CH2 and 5-CH (Chart 2) are
shifting significantly between pH 5.3 and 7.7. The influence
of deprotonation of the tertiary amine on the 3′-CH2 protons
is clear; however, the reason for the change in 5-CH
resonance indicates that the structure of the molecule changes
significantly. A careful analysis of shifts in proton and carbon
resonances corroborates the hypothesis that the molecular
solution structure of TREN-1,2-HOPO changes upon depro-
tonation by 180° rearrangement around the bond between
2′-CH2 and the amide nitrogen (Scheme 1) to maximize
stabilizing hydrogen bond interactions.
Ligand Protonation. Ligand protonation constants were
determined by potentiometry in 0.1 M KCl (Table 1).
Spectrophotometric and NMR titrations confirm these pKa
values (Table S5, Supporting Information). At physiological
pH the ligand is almost completely deprotonated and
available for metal complexation (Figure S2, Supporting
Information). The protonation sites have been assigned based
on the molar absorbances of the species fitting spectropho-
tometric titrations (Figure 3). The very similar absorbances
of the two most basic species (HL2- and L3-) indicate that
the most basic protonation constant pKa 7.16 (1) should be
assigned to the tertiary amine, since this protonation site is
not directly linked to the 1,2-HOPO chromophores monitored
between 260 and 400 nm. This assignment is supported by
1H and 13C NMR titrations (Figures 4 and S3, Supporting
Information, and Tables S4-S6, Supporting Information).
The average of the three pKa values associated with the
three 1,2-HOPO hydroxyl groups of TREN-1,2-HOPO (pKa av
) 4.52) is slightly more acidic than expected based on similar
compounds such as monoacidic propyl-1,2-HOPO (pKa )
4.92) (Chart 3).27 Also in contrast to 1-Me-3,2-HOPO-TREN
1
In addition, assignment of resonances in H and 13C NMR
(27) Xu, J.; Whisenhunt, D. W., Jr.; Veeck, A. C.; Uhlir, L. C.; Raymond,
spectra based on two-dimensional HMBC and HMQC
K. N. Inorg. Chem. 2003, 42, 2665-2674.
9186 Inorganic Chemistry, Vol. 46, No. 22, 2007

[Gd(TREN-1,2-HOPO)(H2O)2] for MRI
Figure 5. Left: Obtained spectra for the competition batch titration of TREN-1,2-HOPO versus DTPA for Gd(III) complexation at pH 7.4 in 0.1 M KCl
solution. The blue line marks the spectrum of Gd-TREN-1,2-HOPO, the red line is uncomplexed ligand, and between are spectra obtained by addition of up
to 85 equiv of DTPA. Right: double logarithmic plot of concentration ratios (at equilibrium; after 24 h) for determination of the conditional stability
constant pGd7.4 19.3 (2).
Chart 2. TREN-1,2-HOPO and Labeling Scheme Used To Assign the
Chart 3. Protonation Constants of Tertiary Amines (Blue) and
Hydroxyl Groups (Red)
1H (Figure 4) and 13C NMR Spectra
Scheme 1. Structural Changes Associated with the Deprotonation of
Tren-1,2-HOPOa
a The deprotonation of the hydroxyl groups (blue) does not change the
hydrogen-bonding pattern. The loss of hydrogen bonding interaction by
removing the proton from the tertiary amine (red) is compensated by
interaction of the lone pair of the resulting amine with amide protons.
Protonation constants of model compounds Pr-3,2-HOPO
(pKa ) 6.12 (1)), Pr-1,2-HOPO (pKa ) 4.92 (2)), and TREN-
MeSAM (pKa ) 6.01 (7)) support this assignment and
illustrate the difference in the acidity of protonation sites of
TREN-1,2-HOPO and TREN-3,2-HOPO. The interaction
between the protonation sites amplifies effects resulting from
deprotonation at adjacent sites, which change the pKa values.
Thus, the pKa values in both TREN-3,2-HOPO and TREN-
1,2-HOPO are lower for the more acidic and higher for the
more basic moieties compared to reference compounds. The
“undisturbed” tertiary amine in the reference compound has
a pKa 6.01 (7), which is more acidic than the 3,2-HOPO
hydroxyl group but more basic than the 1,2-HOPO hydroxyl
group. This reverse assignment of protonation sites of TREN-
1,2-HOPO (amine most basic pKa ) 7.16) compared to
TREN-3,2-HOPO (amine most acidic pKa ) 4.97) is
important for explaining differences in the lanthanide che-
lation and protonation behavior (vide infra).
Scheme 2. Oxalate Binding of Gd-TREN-1,2-HOPOa
a One oxalate anion bound in a bidentate fashion replaces one water
molecule increasing the coordination number at Gd(III) from 8 to 9. This
illustrates the small difference in energy between the eight- and nine-
coordinate geometries.
(pKa ) 4.97) or the model compound TREN-MeSAM (pKa
) 6.01 (7))28 the tertiary amine is more basic (pKa ) 7.16
(1)). This can be rationalized by the hydrogen bond network,
which links the four protonation sites via the amide hydrogen
atoms and by electrostatic interactions caused by additional
charges at the protonated tertiary amine in TREN-1,2-HOPO.
Stability of the Gadolinium Complex. The low solubility
of [Gd(TREN-1,2-HOPO)(H2O)2] (csat ) 85 µM) prevents
potentiometric titrations for stability assessment, and auto-
(28) Cohen, S. M.; Meyer, M.; Raymond, K. N. J. Am. Chem. Soc. 1998,
120, 6277-6286.
Inorganic Chemistry, Vol. 46, No. 22, 2007 9187

Jocher et al.
mated spectrophotometric titration experiments were pre-
vented by slow protonation kinetics. Differences between
the titration from acidic to basic pH and reverse were
observed between pH 4 and 8. Notably, no protonation is
observed above pH 8 since the tertiary amine represents the
most basic function in the ligand. We assign the tertiary
amine of the TREN scaffold as also the most basic function
in this ligand. The observed slow equilibration we attribute
to slowed rotation in the ligand around the CC bond between
carbamide and the aryl ring, most likely occurring upon
change of the protonation state at the tertiary amine. It has
been shown previously for similar ligands that the resulting
hydrogen bond network for the protonated ligand and the
metal complex results in different conformations.29 The fact
that this kinetic hindrance is not observed for TREN-3,2-
HOPO and similar more basic ligands corroborates the
assumption that the amine protonation is crucial for this
behavior.
Figure 6. Comparison of conditional stability constant (pGd) of TREN-
1,2-HOPO (blue) with DTPA (red) and TREN-3,2-HOPO (green) at variable
pH.
Table 2. Association Constants for the Formation of the Ternary
Complexes of Oxalate and a Bidentate 3,2-HOPO Ligand with
Ln-TREN-1,2-HOPO (Ln ) Gd, Eu)
Slow kinetics were circumvented by using batch titration
techniques, which allow for long equilibration times typically
between 1 and 3 days. Competition batch titrations versus
DTPA were employed to determine the conditional stability
constants at pH 6.1 and 7.4 (Figure 5). From these data and
the protonation constants of the ligand, the Gd(III) complex
formation constant logâ110 18.5 (2) was calculated. This
number and the protonation of the complex at pKa 5.0 (2)
and 3.5 (5) are reflected in changes of fluorescence intensity
versus pH of the Eu(III) complex (Figure S10, Supporting
Information) and relaxivity versus pH of the Gd(III) complex
(Figure S11, Supporting Information). In the alkaline region
TREN-1,2-HOPO competes with the formation of europium
hydroxide.30 The competition of oxide formation at high pH
is accurately modeled using logâ110 18.0 (5) and above
europium hydroxide formation constants. In addition the pH
profile of relaxivity suggests that Gd(III) remains coordinated
to very acidic pH values, around 1. This is in agreement
with the suggested stability and the acidic nature of TREN-
1,2-HOPO. Consequently, a non- or only very weakly
luminescent species is formed upon first complex protona-
tion. The increase in relaxivity from pH 4-2 is caused by
decomplexation of one ligand arm from a doubly protonated
complex. While the site of the first complex protonation is
attributed to the scaffold amine, which is the most basic
function in the complex, the second protonation removes one
of the 1,2-HOPO groups leaving space for two additional
water molecules. The increase in relaxivity from 9.5 to 17
mM-1 s-1 reflects this doubling of water coordination sites.
The pGd7.4 19.3 (2) of TREN-1,2-HOPO is remarkably
high considering the acidity of this ligand. A comparison of
conditional stability constants pGd of TREN-3,2-HOPO,
DTPA, and TREN-1,2-HOPO versus pH demonstrates the
advantage of TREN-1,2-HOPO (Figure 6). pGd reaches an
optimum value at pH values at which ligand deprotonation
no longer plays a role in the complexation equilibrium. Thus,
for basic ligands such as DTPA or TREN-3,2-HOPO,
oxalate
oxalate
luminescence
3,2-HOPO
relaxometrya
relaxometrya
logKA (M-1
r1pbound (mM-1 s-1
r1pfree (mM-1 s-1
)
1.90 (5)
5.5
9.3
2.0
3.5
5.2
9.3
)
)
a 20 MHz; 25 °C.
deprotonation still limits the conditional stability constant
pGd at pH 7.4 compared to logâ. By contrast, the more acidic
TREN-1,2-HOPO reaches the “plateau” resulting from
complete ligand deprotonation around physiological pH.
Importantly, for application in MRI pH values above 7.4 play
only a minor role. In contrast, acidotic patients with renal
failure undergoing imaging have been reported to acquire
symptoms of Gd(III) poisoning.31 A more acidic ligand such
as TREN-1,2-HOPO clearly provides increased stability in
the region between pH 1 and 7.
Selectivity for Gd(III) versus Zn(II) and Ca(II). The
toxicity of a MRI-CA correlates not only with its thermo-
dynamic complex stability but also with metal binding
selectivity of the ligand. Thus, the high gadolinium complex
stability of a TREN-1,2-HOPO needs to be completed by
either very slow kinetics for metal exchange or preferably
by maximized thermodynamic selectivity for Gd(III) over
endogenous metals to prevent Gd(III) release in vivo. Zn-
(II) and Ca(II) are the most abundant metals in this respect,
since the concentrations of uncomplexed Zn(II) (10-15 µM)
and Ca(II) (0.4-2.5 mM) in serum are sufficient for release
of lethal amounts of Gd(III) (∼20 µM).1 The selectivity of
TREN-1,2-HOPO for Gd(III) over Zn(II) and Ca(II) was
determined by comparing the conditional stability constants
pGd and pZn or pCa with those for the benchmark compound
DTPA (pGd-pZn ) 4.2; pGd-pCa ) 11.7).
Competition batch titrations against DTPA were employed
to obtain complex stabilities for Zn(II)-TREN-1,2-HOPO
(Table 1, Figure 7). The competition experiments were
(29) Cohen, S. M.; Raymond, K. N. Inorg. Chem. 2000, 39, 3624-3631.
(30) Baes, C. F., Jr.; Mesmer, R. E. The Hydrolysis of Cations; Krieger
Publishing: 1976.
(31) Boyd, A. S.; Zic, J. A.; Abraham, J. L. J. Am. Acad. Dermatol. 2007,
56, 27-30.
9188 Inorganic Chemistry, Vol. 46, No. 22, 2007

[Gd(TREN-1,2-HOPO)(H2O)2] for MRI
Figure 7. Left: obtained spectra for the competition batch titration of TREN-1,2-HOPO versus DTPA for Zn(II) complexation at pH 7.4 in 0.1 M KCl
solution. The blue line marks the spectrum of Zn-TREN-1,2-HOPO and the red line Zn-TREN-1,2-HOPO with 20 equiv of DTPA. Right: double logarithmic
plot of concentration ratios (at equilibrium; after 24 h) for determination of the conditional stability constant pZn7.4 15.2 (2).
performed at pH 6.1 and 7.4. Contrary to the Gd(III) chelate,
no protonation occurs between pH 4.5 and 9. From these
competition experiments pZn 15.2 (2) and logâ110 14.4 (1)
were calculated. The Zn(II) binding of TREN-1,2-HOPO is
relatively strong compared to bidentate 1,2-hydroxypyridi-
none, which forms a quaternary complex ZnL3 with logâ130
11.99.32 This increase in stability for TREN-1,2-HOPO may
be attributed to involvement of the amine nitrogen atom in
binding Zn(II), which also might account for the lack of
observed amine protonation. The selectivity of TREN-1,2-
HOPO for Gd(III) chelation over Zn(II) measured by the
difference in conditional stability constants pGd-pZn 4.1 is
sufficient compared to DTPA with pGd-pZn 4.2.
Figure 8. Observed changes in luminescence of a 10 µM solution of [Eu-
(TREN-1,2-HOPO)] at pH 6 upon titration with concentrated solutions of
several biologically relevant anions.
Competition batch titrations of TREN-1,2-HOPO versus
DTPA for Ca(II) resulted in pCa7.4 8.8 (3) (Figure S9,
Supporting Information). While more stable than Ca-DTPA,
this low stability allows formation of Ca-TREN-1,2-HOPO
only above pH 4 in the presence of physiological amounts
of Ca(II) (2.5 mM). The selectivity of TREN-1,2-HOPO
versus Ca(II) pCa-pGd 10.5 is excellent.
3,2-HOPO ligands that form anionic and cationic Gd(III)
chelates are highly selective for Gd(III) over Ca(II) and Zn-
(II) with ∆pM (Gd-Zn) between 6 and 8.33 Comparing basic
3,2-HOPO with acidic 1,2-HOPO ligands, Zn(II) binding
increases for more acidic ligands, while Gd(III) binding
decreases for such ligands. Thus, for the design of highly
selective and acid resistant MRI-CA, mixed ligands of acidic
1,2-HOPO and more basic 3,2-HOPO groups are suggested.
Alternately, the number of acidic chelate groups could be
raised to four resulting in an octadentate ligand with
preferential binding of lanthanides.1
plexes. In addition solid-state crystal structures of ternary
complexes were obtained to correlate solid and proposed
solution structures.34 Binding constants of phosphate and
oxalate are relatively low for DOTA and DTPA, but cationic
complexes of DO3A amide derivatives (q ) 2) can efficiently
bind bidentate carboxylates such as citrate or malonate.35
Unlike these cationic chelates, neutral 3,2-HOPO complexes
with q ) 2 do not or only weakly bind biorelevant anions.
Also a cationic and an anionic bis-3,2-HOPO-TAM chelate
show smaller association constants with phosphate and
oxalate than the commercial, q ) 1 chelates Gd(DOTA) or
Gd(DTPA).33
The affinity of Ln-TREN-1,2-HOPO toward several anions
was evaluated by relaxometry of the Gd(III) complex and
luminescence spectrosocopy of the Eu(III) complex (Table
2). Endogenous anions like phosphate, carbonate, acetate,
or malonate do not interact with both complexes (Figure 8).
A weak interaction with oxalate was detected by decreasing
the relaxation rate (Figure 9) and increasing luminescence
intensity. For complete formation of the ternary complex
Anion Affinity. Anion binding at the water coordination
site reduces the efficiency of a MRI-CA by replacing these
water molecules. For aminocarboxylate ligands, anion bind-
ing has been studied with a chiral macrocycle by lumines-
cence spectroscopy of the Eu(III) complex, relaxometry of
Gd(III) chelates, and NMR spectroscopy of Yb(III) com-
(34) Dickins, R. S.; Aime, S.; Batsanov, A. S.; Beeby, A.; Botta, M.; Bruce,
J. I.; Howard, J. A. K.; Love, C. S.; Parker, D.; Peacock, R. D.;
Puschmann, H. J. Am. Chem. Soc. 2003, 124, 12697-12705.
(35) Botta, M.; Aime, S.; Barge, A.; Bobba, G.; Dickins, R. S.; Parker,
D.; Terreno, E. Chem.--Eur. J. 2003, 9, 2102-2109.
(32) Li, Y. J.; Martell, A. E. Inorg. Chim. Acta 1993, 214, 103-111.
(33) Pierre, V. C.; Botta, M.; Aime, S.; Raymond, K. N. Inorg. Chem.
2006, 45, 8355-8364.
Inorganic Chemistry, Vol. 46, No. 22, 2007 9189

Jocher et al.
Figure 9. Relaxometric determination of the association constants KA for oxalate (left) and a bidentate 3,2-HOPO ligand (right) with Gd-TREN-1,2-HOPO
(c ) 35 µmol; 20 MHz; 25 °C).
1500 equiv of oxalate are required, which corresponds to a
very small binding constant of Ka 82.7 ( 6.5 M-1 for the
Gd(III) complex and 96 M-1 for the Eu(III) complex.
Ascorbate and catecholate decrease the luminescence
intensity of the Eu(III) complex, but no interaction was
observed by relaxometry indicating the mechanism for
luminescence quenching of the Eu(III) complex does not
involve coordination at the metal center. It is possible the
ligand triplet state may be quenched by catechol or ascorbate,
or, alternately, the excited Eu(III) may be partially reduced
to Eu(II). The excited state of Eu(II),36 however, is too high
in energy for sensitization by TREN-1,2-HOPO.
It was also noted that the addition of up to 1000 equiv of
fluoride did not replace water. The only strong binding
interaction was observed for a bidentate 3,2-HOPO anion
(Figure 9). The binding constant is within the expected value
around 3100 M-1 (logKa 3.5). This observation corroborates
Figure 10. Observed emission spectrum for the oxalate (c ) 15 mM)
adduct of [Eu(TREN-1,2-HOPO)(H2O)2] (c ) 10 µM) in aqueous solution
the assumption that only monoanionic and bidentate ligands
would form strong ternary complexes with Eu-TREN-1,2-
HOPO. The 3,2-HOPO anion binds 1.5 order of magnitude
stronger than oxalate, with saturation already below 20 equiv,
while other bidentate ligands like catechol or maltol do not
bind.
Solution Structure of Ternary Complexes. The change
of relaxation rates between Gd-TREN-1,2-HOPO and the
ternary adducts suggests that the coordination number
increases from 8 to 9 upon binding of bidentate 3,2-HOPO
or oxalate (Figure 10). Indeed, the calculated relaxivities r1p
(20 MHz, 25 °C) of the ternary complexes (Table 2) are 5.5
( 0.2 and 5.2 ( 0.2 mM-1 s-1. These are typical values for
q ) 1 complexes. Pure outer sphere complexes (q ) 0) would
be characterized by relaxivity values of ca. 2.5-3.0 mM-1
s-1 under identical experimental conditions.
5
at pH 6. The inset shows the expansion of emission spectra in the D0 f
7FJ (J ) 0, 1; 570-600 nm) region and corresponding spectral deconvolution
into several overlapping Voigt functions.
ternary complex is the dominant species, which is also
recognized by slight changes in the emission spectrum.
Lifetime measurements clearly indicated the new species has
one coordinated water (q ) 0.8).24 The fact that ligands with
one carboxylate group do not coordinate under these
experimental conditions corroborates that oxalate functions
as a bidentate ligand replacing only one water molecule. Such
behavior was observed earlier in the oxalate binding of a
tren substituted bis-3,2,-HOPO-TAM-TREN complex, which
was interpreted as shifting the equilibrium from eight- to
nine-coordinate complexes upon addition of oxalate.33 The
emission spectrum of a 10 µM Eu-TREN-1,2-HOPO solution
with 1500 equiv of oxalate differs slightly from the one of
a pure Eu-TREN-1,2-HOPO solution. Symmetry analysis of
the J ) 0 and 1 bands (Figure 10) for a nine-coordinate Eu-
(III) center suggests that the coordination geometry is a
subgroup of C4V, since otherwise the J ) 0 band would be
absent (Table S3, Supporting Information). Fitting the J )
1 band with three equally broad and intense Voigt functions
Titrations of the Eu(III) complex with oxalate were
monitored by variation of luminescence decay lifetimes
which rise from 220 µs to 332 µs upon addition of 1500
equiv of potassium oxalate. Under these conditions the
(36) Adachi, G.; Tomokiyo, K.; Sorita, K.; Shiokawa, J. J. Chem. Soc.,
Chem. Commun. 1980, 914 - 915.
9190 Inorganic Chemistry, Vol. 46, No. 22, 2007

[Gd(TREN-1,2-HOPO)(H2O)2] for MRI
(A and 2B) provides a better result than with two Voigt
functions of a 2:1 intensity ratio (A and E). Thus the
coordination geometry at the metal center can be assigned
to C2, a subgroup of C4V.
Despite the differences, 1,2-HOPO chelates overall serve
as good models for nonluminescent 3,2-HOPO chelates. The
higher acidity of the 1,2-HOPO chelate group compared to
3,2-HOPO facilitates handling and analysis since the com-
plexes are less sensitive to acid hydrolysis. The advantages
of using 1,2-HOPO over 3,2-HOPO are the lower anion
affinity, higher acid resistance of the lanthanide chelate, and
the utilization of Eu(III) luminescence to reveal information
on the solution structure. Further research aims on improving
the water solubility of 1,2-HOPO derivatives.37
Conclusion
The X-ray structures of [Eu(TREN-1,2-HOPO)(H2O)2] and
[Gd(TREN-1,2-HOPO)(H2O)2] are isostructural with identi-
cal coordination geometries at the metal centers. This allows
the direct comparison of relaxometric measurements of the
Gd(III) complex with the emission spectra of the Eu(III)
complex. The protonation constants of TREN-1,2-HOPO
were determined and assigned revealing a change in the order
of protonation sites, which also influences the sequence of
protonating the Gd(III) complex. With the more acidic 1,2-
HOPO hydroxyl groups, the tertiary amine becomes the most
basic function contrary to previously studied more basic 3,2-
HOPO ligands. The high Gd(III) binding constant of TREN-
1,2-HOPO is matched by sufficient selectivity versus Zn(II)
and excellent selectivity versus Ca(II) binding. In addition,
the greater ligand acidity provides enhanced stability under
slightly acidic conditions. The expectation of low toxicity
for Gd(III)-TREN-1,2-HOPO based on metal selectivity and
complex stability is completed by low anion interactions of
the complex to maintain relaxivity in vivo. A detailed
study on the ternary oxalate adduct reveals insights into
solution structure and demonstrates the comparable energies
of the -eight- and nine-coordinate Gd(III), which is an
essential feature for the fast water exchange rates that are
one prerequisite for achieving advantageous relaxometric
properties of the MRI contrast agents based on HOPO
chelates.
Acknowledgment. This work (UCB) was supported by
NIH Grant HL69832, by the Director, Office of Science,
Office of Basic Energy Sciences, and the Division of
Chemical Sciences, Geosciences, and Biosciences of the U.S.
Department of Energy at LBNL under contract DE-AC02-
05CH11231 and NATO travel grant PST.CLG.980380. C.J.J.
acknowledges the German Research Foundation (DFG) for
a postdoctoral fellowship. Support from MIUR (COFIN
2005) is also gratefully acknowledged (M.B., S.A.). The
authors acknowledge Eric Werner for helpful discussions and
proofreading of the manuscript.
Supporting Information Available: Single-crystal X-ray dif-
fraction data (CIF) and additional tables (symmetry analysis and
titration refinements) and figures (2D NMR data, titration data, pH-
luminescence profile, pH relaxivity profile). This material is
available free of charge via the Internet at http://pubs.acs.org.
IC700985J
(37) Seitz, M.; Pluth, M. D.; Raymond, K. N. Inorg. Chem. 2007, 46, 351-
353.
Inorganic Chemistry, Vol. 46, No. 22, 2007 9191