A triazacyclononane-based bifunctional phosphinate ligand for the preparation of multimeric68GA tracers for positron emission tomography

Article abstract of DOI:10.1002/chem.200903281

For application in positron emission tomography (PET), PrP9, a N,N′,N″-trisubstituted triazacyclononane with methyl(2-carboxyethyl) phosphinic acid pendant arms, was developed as ⁶⁸Ga³⁺ complexing agent. The synthesis is short and inexpensive. Ga^III and Fe^III complexes of PrP9 were characterized by single-crystal X-ray diffraction. Stepwise protonation constants and thermodynamic stabilities of metal complexes were determined by potentiometry. The Ga^III complex possesses a high thermodynamic stability (log K_[GaL]=26.24) and a high degree of kinetic inertness. ⁶⁸Ga labeling of PrP9 is possible at ambient temperature and in a wide pH range, also at pH values as low as 1. This means that for the first time, the neat eluate of a TiO₂based ⁶⁸Ge/⁶⁸Ga generator (typically consisting of 0.1M HCl) can be directly used for labeling purposes. The rate of ⁶⁸Ga activity incorporation at pH 3.3 and 20°C is higher than for the established chelators DOTA and NOTA. Tris-amides of PrP9 with amino acid esters were synthesized to act as models for multimeric peptide conjugates. These conjugates exhibit radiolabeling properties similar to those of unsubstituted PrP9.

Full text of DOI:10.1002/chem.200903281

DOI: 10.1002/chem.200903281
A Triazacyclononane-Based Bifunctional Phosphinate Ligand for the
68
Preparation of Multimeric Ga Tracers for Positron Emission Tomography
Johannes Notni,*[a, b] Petr Hermann,*[b] Jana Havlꢀcˇkovꢁ,[b] Jan Kotek,[b]
Vojteˇch Kubꢀcˇek,[b] Jan Plutnar,[b] Natalia Loktionova,[c] Patrick Johannes Riss,[c]
Frank Rçsch,[c] and Ivan Lukesˇ[b]
Abstract: For application in positron
emission tomography (PET), PrP9, a
sesses a high thermodynamic stability
(logK[GaL] =26.24) and a high degree of
kinetic inertness. 68Ga labeling of PrP9
is possible at ambient temperature and
in a wide pH range, also at pH values
as low as 1. This means that for the
first time, the neat eluate of a TiO2-
based 68Ge/68Ga generator (typically
consisting of 0.1m HCl) can be directly
used for labeling purposes. The rate of
68Ga activity incorporation at pH 3.3
and 208C is higher than for the estab-
lished chelators DOTA and NOTA.
Tris-amides of PrP9 with amino acid
esters were synthesized to act as
models for multimeric peptide conju-
gates. These conjugates exhibit radio-
labeling properties similar to those of
unsubstituted PrP9.
N,N’,N’’-trisubstituted
triazacyclono-
nane with methyl(2-carboxyethyl)phos-
phinic acid pendant arms, was devel-
oped as 68Ga3+ complexing agent. The
synthesis is short and inexpensive. GaIII
and FeIII complexes of PrP9 were char-
acterized by single-crystal X-ray dif-
fraction. Stepwise protonation con-
stants and thermodynamic stabilities of
metal complexes were determined by
potentiometry. The GaIII complex pos-
Keywords: gallium
ligands · potentiometry · radiophar-
maceuticals · stability constants
·
macrocyclic
AHCTUNGTRENNUNG
Introduction
or magnetic resonance imaging (MRI). For this technique,
emission of positrons by b+-emitters or, more precisely, the
g-radiation originating from their annihilation with elec-
trons, is detected. At present, PET facilities are still depen-
dent on an on-site cyclotron, since the predominantly used
radioisotopes (18F, 11C, 13N, and 15O) are cyclotron-produced
and have very short half-lives. This renders PET an expen-
sive imaging technique. Recently, the positron emitter 68Ga
has gained more attention as it is now available from com-
mercially distributed radionuclide generators. In such devi-
ces, 68Ga is produced by decay of the parent nuclide 68Ge
(T1/2 =271 d), which is absorbed on a TiO2 or SnO2 matrix.
68Ga-based radiopharmaceuticals are thus available inde-
pendently of an on-site cyclotron and enable medical PET
scans at a fraction of the usual cost. It possesses a half-life
of ꢀ68 min and is thus compatible with most bio-targeting
applications. The topic of utilization of 68Ga in PET has re-
cently been reviewed.[1] In addition, there are two more cy-
clotron-produced gallium isotopes suitable for nuclear imag-
ing: 1) 67Ga is a g-emitter (T1/2 =78.3 h) used in single-
photon emission computed tomography (SPECT) and thera-
py; 2) b+-emitting 66Ga with an intermediate half-life of
Positron emission tomography (PET) is a powerful tool in
non-invasive and quantitative medical imaging, augmenting
other imaging methods such as computer tomography (CT)
[a] Dr. J. Notni
Department of Nuclear Medicine
Technische Universitꢀt Mꢁnchen
Ismaninger Str. 22, 81675 Mꢁnchen (Germany)
Fax : (+49)89-4140-4841
E-mail: johannes.notni@tum.de
ˇ
[b] Dr. J. Notni, Dr. P. Hermann, J. Havlꢂckovꢃ, Dr. J. Kotek,
ˇ
ˇ
Dr. V. Kubꢂcek, Dr. J. Plutnar, Prof. Dr. I. Lukes
Department of Inorganic Chemistry, Faculty of Science
Charles University in Prague, Hlavova 2030
Praha 2, 12840 (Czech Republic)
Fax : (+420)22195-1253
E-mail: petrh@natur.cuni.cz
[c] N. Loktionova, Dr. P. J. Riss, Prof. Dr. F. Rçsch
Institute of Nuclear Chemistry, University Mainz
Fritz-Strassmann-Weg 2, 55128 Mainz (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200903281.
7174
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7174 – 7185

FULL PAPER
9.4 h did not receive much attention in the past,[2] but the in-
terest in this radioisotope has revived recently.[3]
In aqueous solutions, gallium is stable only as a trivalent
cation. It cannot be incorporated into the structure of tar-
geting vectors by covalent bonding, but must be complexed
by a ligand that is conjugated to the biological vector. The
Ga3+ ion possesses a d10 electron configuration and accepts
different coordination numbers (usually 4–6), while not dis-
playing preference for any particular coordination poly-
hedron. At pH>4, formation of colloidal hydroxide
[Ga(OH)3]n commences. Although this does not generally
inhibit complex formation, radiolabeling is nevertheless sub-
stantially hampered due to formation of insoluble colloids
(particularly at high activities) and their adhesion to the sur-
face of the reaction vessel. At pH values above 8, a water-
soluble hydroxo complex, [Ga(OH)4]ꢁ, is formed. As ligand
exchange with the tetrahydroxo complex is a much slower
process than complexation of free Ga3+, complexation is
achieved best at pH<4.
Ligands for 68Ga-based PET radiopharmaceuticals should
ideally combine the following set of properties.
1) Stability: GaIII complexes should be as stable as possible;
a kinetic inertness of the complex is more important
than high thermodynamic stability.
2) Quick complexation under radiochemical conditions:
Formation of GaIII complexes should be fast at low tem-
peratures, low concentration, and minimal excess of the
ligand. A desirable ligand will chelate Ga3+ in solutions
of nanomolar concentration at room temperature within
minutes.
3) Selectivity: The ligand should ideally be selective for
Ga3+ ion. Particularly, complexation of serum metals
like Ca2+, Mg2+, and Zn2+ ions (the last being produced
by decay of 68Ga) should be disfavored in order to avoid
transmetallation in vivo or diminishing of radiochemical
yield.
4) Conjugation ability: The chelating unit has to possess a
functional group which allows covalent binding to the
targeting vector (biomolecule) without a significant dero-
gation of complexation performance.
5) Long shelf life: In medical applications, excellent chemi-
cal stability is necessary.
are less rigid and do not have a cavity of a particular size,
which is why they are less selective for particular ions, but
show faster complex formation.[6,7] Despite of the amazing
thermodynamic stabilities of the resulting complexes,[8] li-
gands containing thiol groups are disfavored because of
their tendency to degrade, albeit slowly, by oxidation. Also,
the comparably low acidity of sulfhydryl groups hampers
complex formation in acidic media, and derivatives suitable
for bioconjugation are difficult to synthesize.[9]
DOTA derivatives are currently the working horses of
bioACTHNUGRTENUNGconjugational chemistry related to medical imaging, in-
cluding 68Ga applications.[10,11] This is mainly due to the com-
mercial availability of ready-to-use mono-unprotected pre-
cursors like DOTA
fact that lanthanideACTHNUGTRNE(GUN III) complexes of DOTA derivatives
(tBu)3,[12] which, in turn, is rooted in the
ACHTUNGTRENNUNG
and its conjugates have been extensively used as MRI con-
trast agents[13–16] and radiotherapeutics.[17] However, NOTA-
like ligands show a much better selectivity towards Ga3+ ion
and their complexes are more stable, as the size of the
NOTA cavity is almost ideal for this ion.[18–20] Conjugable
NOTA derivatives like NODASA,[21] NODAGA,[22]
NODAPA-NCS[23] as well as other thiocyanato-equipped
NOTAs[24,25] are thus much better suited for the synthesis of
68Ga radiopharmaceuticals.
With all these considerations in mind, we devised a ligand
structure that combines the advantages of the different
structural motifs. The novel chelator PrP9 (shown here) is
structurally related to some known [9]aneN3 derivatives.[26–29]
Unlike the above-mentioned ligands, it contains phosphi-
nates instead of carboxylates as primary coordination sites.
Due to the lower pKa value of phosphinic acids (mostly <1)
compared to aliphatic carboxylic acids (typically ꢀ4),[30]
6) Accessibility: Preparation of the compound in practical
amounts should be quick, facile, and inexpensive.
A choice of ligands (some of them representing a family
of compounds) which have been proposed or have already
been used for application in 68Ga radiopharmaceuticals is
depicted here (see references [8,12,18,19,21,23,27,58–64]).
All of them have advantages and drawbacks. Aza-macrocy-
cle based ligands generally form complexes with increased
thermodynamic and, more importantly, kinetic stability com-
pared to open-chain or tripod ligands.[4,5] However, in case
of the first group, a considerable barrier has to be overcome
in order to place the metal ion into the ligand cavity, thus
causing slow complexation kinetics. The open-chain ligands
Chem. Eur. J. 2010, 16, 7174 – 7185
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7175

J. Notni, P. Hermann et al.
Scheme 1. Synthesis of the hexaprotonated (zwitterionic) form of the
ligand PrP9.
metal-ion complexation should be possible at much lower
pH values than in case of NOTA and DOTA, thereby ex-
panding the pH range of application. Furthermore, the dis-
tant carboxylates are supposed to act as “pre-coordination”
sites: at very low concentrations typical for radiochemistry,
fast, open-chain-like interactions with metal ions are sup-
posed to help to increase the effective metal concentration
close to the ligand cavity and therefore increase complexa-
tion rate. The structure is thus expected to combine advan-
tages of both macrocycle-based chelators (stability and se-
lectivity) and open-chain ligands (fast complex formation).
In addition, phosphinic acid and carboxylic groups exhibit
different reactivity, which can be utilized to functionalize
the carboxylates without having to protect the phosphinate
moieties. The availability of three carboxylic acid moieties
for conjugation furthermore paves the way towards multi-
meric tracers.
Complex synthesis and solid-state structures: Although
PrP9 forms complexes with a variety of metal ions, precipi-
tates suitable for X-ray diffraction and analysis could only
be obtained for [GaACTHNUTRGENN(UG H3PrP9)] and [FeACHTUNTGERN(NUGN H3PrP9)]. Two
phases were obtained for each, differing only in the amount
of water of crystallization. From acetone/water mixtures, the
complexes crystallized as dihydrate and monohydrate, re-
spectively, whereas a simple evaporation of an aqueous solu-
tion afforded isostructural hexahydrates. In both compounds
the chelate cages in the two phases are quite similar and just
one example is depicted (Figure 1; for the FeIII complex see
Results and Discussion
Throughout this paper the abbreviation PrP9 (systematic
name of PrP9 is 1,4,7-triazacyclononane-1,4,7-trisACTHNUTRGNEUG[N methyl(2-
carboxyethyl)phosphinic acid]) is used regardless of proto-
nation state, except in cases in which the distinction is neces-
sary for comprehension.
Ligand synthesis: The preparation protocol for PrP9 is ex-
tremely short; there are just three steps starting from stock
chemicals and only the last reaction involves the azamacro-
cycle. All steps have satisfying to excellent yields. The com-
plete synthesis is outlined in Scheme 1 and can be carried
out in less than three days. All required materials are com-
mercially available and, apart from the amine [9]aneN3, they
are also inexpensive. In addition, the synthesis requires very
little workup effort. One extraction after the first step and
just a simple ion exchange chromatography after the third
step, followed by recrystallization, are necessary in order to
obtain a very pure product. The procedure can be scaled up
without problems, as has been proven by preparing 15 g of
PrP9 in a single batch. For these reasons, we hold the view
that PrP9 has good prospects in providing a basis for the
synthesis of bioconjugated chelators.
Figure 1. Molecular structure of the [Ga
AHCTUNGTRENNUNG
ACHTUNGTREN(NUGN H3PrP9)] complex in [Ga-
Supporting Information, Figure S1). Coordination bond
lengths and angles of the two GaIII structures are almost
identical, whereas the two FeIII phases exhibit small structur-
al differences (see Table 1). The parameters of the [Ga-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
in a trigonal-antiprismatic coordination sphere and the pro-
pionic acid moieties being located above the complex cage.
An identical structural arrangement has been found in the
7176
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7174 – 7185

Radiopharmaceuticals
FULL PAPER
Table 1. Selected parameters of the complex moieties in [Fe-
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG(H3PrP9)]·nH2O and [GaCAHTUNGTRENNNUG
given in ꢅ, angles in 8].
A
ACHTUNGTRENNUNG
ꢁ
ꢁ
The Fe O and Ga O distances in [Ga
AHCTUNTGRENGN(UN H3PrP9)] (see Table 1) are almost identical, but Fe N
ACHTUNGTREN(UNNG H3PrP9)] and (Fe-
n=1
n=6
n=2
n=6
ꢁ
ꢁ
M N1
ꢁ
M N4
2.210(2)
2.232(2)
2.189(2)
1.927(1)
1.932(1)
1.934(1)
162.27(5)
162.19(5)
164.31(5)
48.41
47.79
51.01
2.3929
1.487(2)
0.9063(2)
3.57(8)
2.178(2)
2.198(2)
2.178(2)
1.948(1)
1.947(1)
1.946(1)
164.54(6)
164.73(6)
165.41(6)
49.91
49.44
51.37
2.3966
1.4540(2)
0.9428(2)
1.63(6)
2.120(2)
2.124(2)
2.125(2)
1.926(1)
1.924(1)
1.929(1)
169.64(6)
169.63(6)
169.95(6)
52.08
51.90
52.78
2.3722
1.3594(2)
1.0128(2)
0.51(3)
2.108(1)
2.128(1)
2.106(1)
1.937(1)
1.931(1)
1.932(1)
170.07(5)
169.88(5)
170.21(4)
51.82
51.55
53.16
2.3788
1.3506(2)
1.0284(2)
1.22(5)
ꢁ
bonds are significantly longer than Ga N linkages. Also, the
coordination environment of FeIII is more distorted, the
average N-M-O and the torsion angles of the GaIII com-
plexes being closer to the ideal values. This means that the
cavity size of PrP9 is slightly less suitable for FeIII (ionic
radius 0.69 ꢅ) than for GaIII (ionic radius 0.76 ꢅ).
ꢁ
M N7
ꢁ
M O11
ꢁ
M O21
ꢁ
M O31
N1-M-O31
N4-M-O11
N7-M-O21
N1-NQ-OQ-O11[a,b]
N4-NQ-OQ-O21[a,b]
Solution structure of [GaACTHGNUTERNNUG
(PrP9)]: In both 31P and 71Ga
NMR spectra, single resonances were found (42.4 and
ꢀ135 ppm, respectively). The 31P NMR thus proves that
solid-state and solution structures are identical, that is, only
one non-fluxional pair of enantiomers Lddd-RRR/Dlll-SSS
is present. The half-width of the Ga resonance (n1/2
N7-NQ-OQ-O31[a,b]
[a]
ꢁ
NQ OQ
[a]
ꢁ
M NQ
[a]
ꢁ
M OQ
N3/O3 planes ]
ꢀ400 Hz) is larger than in case of [Ga
(NOTA)] (210 Hz)[19]
ACHTUNGTRENNUNG
[a] NQ and OQ are the barycenters of the N3 and O3 planes, respectively.
[b] Torsion angle.
and more close to the values found for complexes of related
phosphorus containing ligands, the methylenephosphonate
analogue [Ga
phosphinate analogue [Ga
(NOTP)]3ꢁ (430 Hz)[35] or methylene
ACHTUNGTRENGNUN ACHTUNRTGENN(UG phenACHTUNGTRENNUNGyl)-
G
ACHTUNGTRENNUNG
1
[Ga
late-ring conformation and pendant-arm helicity has been
found in the structurally related compounds [Ga
(NOTA)][19]
and [Ga
(NODASA)][21] (Table 2), as well as in other GaIII
ACHTUNGTRENNUNG
(NOTPPh)] complex.[32] A similar combination of che-
lets at 2.16 and 2.76 ppm, originating from protons of the
propionate side arms, can be unambiguously assigned. The
resonances of the remaining protons form a set of overlap-
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
ping multiplets, indicating
a
rigid structure of the whole
complex, which appears to be
Table 2. Comparison of some structural parameters of [GaACTHNUTRGNEUNG(H3PrP9)] with analogous complexes [distances
given in ꢅ, angles in 8].
conformationally
stable
as
1H NMR spectra measured at
temperatures of 20–808C indi-
cate a high rigidity of the ligand
skeleton. Up to 608C, the split-
ting pattern of the ring protons
and the nitrogen-bound side-
arm protons (N-CH2-P) remains
unchanged; some signal coales-
cence starts at 808C. This is in
N-Ga-O[a]
Torsion ][b]
OQ-NQ[c]
Ga-OQ[c]
Ga-NQ[c]
[Ga
[Ga
[Ga
[Ga
N
169.9
168.3
167.4
165.5
52.2
52.1
47.6
44.4
2.376
2.370
2.329
2.363
1.355
1.391
1.318
1.333
1.021
0.979
1.012
1.030
[a] Average of values for opposing N and O atoms. [b] For definition of torsion angle see Table 1. [c] NQ and
OQ are the barycenters of the N3 and O3 planes, respectively. [d] Average values of two structures.
complexes with six-membered pendant arm chelate
rings.[33,34] Interestingly, a different situation is encountered
for FeIII complexes: for phosphorus-containing ligands,
pendant arm helicity is opposite to macrocycle chelate rings
(Lddd/Dlll, similarly as found in the Ga structures); for lig-
ACHTUNGTRENNUNGands with carboxylate pendant arms it is the same (Llll/
Dddd; for more information see the Supporting Informa-
tion).
accordance with the respective observations made for [Ga-
(NOTA)], [Ga(NOTPPh)], and [Ga
(NOTP)]3ꢁ com-
plexes.[19,32,35]
G
E
ACHTUNGTRENNUNG
In the 13C NMR spectrum, the signals of the ring carbons
(appearing as a single resonance in the solution of the free
ligand) are split into two separate signals, one of them show-
ing the C,P coupling (11 Hz, Figure S4). This further corrob-
orates the conclusions drawn from 31P NMR spectra. No co-
alescence of these signals was observed at temperatures up
to 908C. Hence, the lower limit for the interconversion
Gibbs free energy between the two enantiomers can be ap-
proximately assessed to DGint >100 kJmolꢁ1.
Comparison of average “diagonal” N-Ga-O and torsion
angles of structurally related GaIII complexes (Table 2) re-
veals that the values of those derived from phosphinate lig-
ACHTUNGTRENNUNGands, and particularly of [GaACHTUGNTREN(NUNG H3PrP9)], are more close to
those of an ideal octahedron (1808 and 608, respectively).
Also, the N3 and O3 planes are slightly more rotated. In all
structures regarded, gallium is located closer to the N3 than
to the O3 plane; the difference is most pronounced in the
structure of the phenylphosphinate complex [Ga-
Thermodynamic and mechanistic studies: Potentiometry and
NMR techniques were employed for determination of pro-
tonation constants and complex stabilities as well as for in-
vestigation of the mechanism of complex formation.
Chem. Eur. J. 2010, 16, 7174 – 7185
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7177

J. Notni, P. Hermann et al.
The stepwise protonation constants of the ligand
(Figure 2) are logKa =11.48, 5.44, 4.84, 4.23, 3.45, and 1.66.
The protonation sequence can be estimated by a comparison
ion. Complexes with Cu2+ and Zn2+ are about ten, and
those with Mg2+ and Ca2+ approximately 20 orders of mag-
nitude less stable. Furthermore, stability constant values for
the related metal-ion pairs Mg2+/Ca2+ and La3+/Gd3+ illus-
trate the ligandꢆs selectivity for small ions, as for both cou-
ples the stability constant of the larger ion is about two
orders of magnitude lower. Similar selectivities have been
observed for NOTA[37,40] and its phosphinate analogues.[27,28]
(For a comparison of stability constants and more distribu-
tion diagrams see Supporting Information; Table S4, Fig-
AHCTUNGTREGuNNNU res S2 and S3). Also, PrP9 fits into the correlation of over-
all basicity of these ligands with the stability constants of
their CuII and ZnII complexes.[30]
The distribution diagram (Figure 3) shows that the 1:1
complex is triply negatively charged under physiological pH.
In solutions containing two equivalents of Ga3+, the primary
complex appears to be able to bind an additional Ga3+ ion,
most likely by weak interaction with the side-arm carbox
ACHTUNGTRENNUNGyl-
AHCTUNGTRENNUNG
form of an “out-of-cage” complex). This shows the ability of
the carboxylates to direct Ga3+ ions towards the ligand
cavity, which we believe is the main cause for the unusually
fast complex formation.
Figure 2. Distribution diagram for the protonation of PrP9.
of these data with those for similar ligands (see also Sup-
porting Information, Table S2).[27,28,36] The first two protona-
tions should occur on the ring nitrogen atoms, followed by
three protonations on the side-arm carboxylates. The other
protons are most likely bound on the remaining ring nitro-
gen atom and/or phosphinate groups. Thus, not all phosphi-
nate groups are protonated, even in solutions with a very
low pH (<1) and therefore are still able to complex Ga3+
ion. Basicity of the ring nitrogen atoms is somewhat lower
than in case of NOTA,[37] but higher than for triazacyclono-
nane-derived methylphosphinic acid ligands,[27,28] presumably
due to higher overall negative charges of the PrP9 anions.
Basicity of the distant carboxylate groups is in the usual
range for carboxylic acids.[38,39]
The stability constants were determined for complexes of
PrP9 with several biologically important metal ions as well
as some lanthanides. For Y3+ and Lu3+ ions, the measure-
ments were not possible, as insoluble precipitates were ob-
tained, most likely due to formation of complex polymers
containing protonated ligand molecules. The data given in
Table 3 clearly show the preference of PrP9 for the Ga3+
From Figure 3 it is furthermore apparent that the GaIII
complex is fully formed, even at the beginning of titration,
in the form of a fourfold protonated species, [GUNNENTRGUNHTCA
Ga(H4PrP9)]+,
the corresponding fourth protonation constant being logKa
ꢀ0.7 (Table 3). Hence, the K[GaL] value obtained from po-
tentiometry is determined by the competitive ligand ex-
change reaction with an hydroxide anion, which, in order to
represent the complexesꢆ actual thermodynamic stability,
must be in thermodynamic equilibrium. This reaction, how-
ever, turned out to proceed surprisingly slowly and this had
to be taken into account in order to obtain correct results.
During the normal “in-cell” titration, we observed no pre-
cipitate of Ga(OH)3 and pH readout quickly stabilized over
the whole pH range of measurement (1.5–12). From the
data thus collected, a stability constant of logK[GaL] =35.65
was calculated. However, this value appeared unrealistically
high compared to the one of the [GaACHTNUTRGNEUNG(NOTA)] complex
(logK[GaL] =31.0),[18] since PrP9 exhibits a lower overall ba-
sicity than NOTA. When the reaction was given a timespan
of about four weeks in order to reach equilibrium (“out-of-
cell” titration), we obtained a
more
realistic
value
of
logK[GaL] =26.24, which is, as
Table 3. Stability constants [logK] and stepwise protonation constants [logKA] of complexes of PrP9 with se-
lected metal ions.
expected, lower than that of
[GaACTHNUTRGNE(NUG NOTA)]. Even at pH 11,
Ga3+
Cu2+
Zn2+
Mg2+
Ca2+
La3+
Gd3+
the reaction half-life was ꢀ60 h
(see Supporting Information,
Figure S7); finally, at pH 13 the
reaction was completed within
minutes. A possible explanation
is the high negative charge of
L+MÐLM
26.24
5.2
4.5
3.8
0.7
16.85
5.14
4.66
3.95
1.33
12.24
16.88
5.17
4.68
3.96
7.84
6.49
5.00
4.74
4.29
6.04
7.94
4.98
4.50
11.26
6.22
5.00
3.95
3.32
13.46
4.80
4.80
3.78
3.38
LM+HÐHLM
HLM+HÐH2LM
H2LM+HÐH3LM
H3LM+HÐH4LM
LM(OH)+HÐLM
N
9.9
12.63
13.04
2.80
12.8
11.32
12.30
10.42
11.23
LM(OH)2 +HÐLM(OH)
ACHTUNGTRENNUNG
the [GaACHTUNGTRNEUNG
(PrP9)]3ꢁ ion, which
LM+MÐLM2
7.3
2.5
3.32
4.79
2.43
4.71
2.95
5.9
8.5
LM2 +HÐHLM2
LM2(OH)+HÐLM
could be hampering the ap-
(H2O)
proach of OHꢁ to the central
7178
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7174 – 7185

Radiopharmaceuticals
FULL PAPER
libration requires weeks which renders it irrelevant in terms
of 68Ga radiochemistry.
The complexation mechanism was also investigated by 31P
and 71Ga NMR spectroscopy. Between pH 1.5 and 8, quanti-
tative complexation occurs immediately after mixing equi-
molar amounts of Ga3+ ions and PrP9 without any inter-
mediates being detectable. However, at initial pH values of
1.3, 1.0, and 0.8, quantitative complexation requires 8, 65,
and 90 min, respectively (Figure 4 and Figures S6 and S7 in
Figure 3. Distribution diagrams for the GaIII/PrP9 systems (ligand-to-
metal-ratio: upper: L/M=1:1, lower: L/M=1:2; cL =4 mm).
ion.[18] This shows that neglecting the kinetics of the ligand–
hydroxide competition for the galliumACTHNUTRGNE(NUG III) ion can easily
lead to erroneous values for thermodynamic stabilities; such
kinetic inertness may give a false impression of a higher sta-
bility constant. For polydentate/rigid ligands, equilibration
time generally should be checked carefully, even in the alka-
line pH region at which, until now, equiCAHTUNGTRENlUNNG ibACHTUNGTREGrNNUN iACTHNUGTRENuNNGU ms have been
Figure 4. Complexation of Ga3+ with PrP9 as followed by 31P (top) and
71Ga NMR (bottom) (M/L=1, pH 0.8, 258C). Peak labeling: *: final (“in
supposed to establish quickly. The carboxylate protonation
constants for the GaIII complex obtained by both “non-equi-
librium” (logKa =5.14, 4.54, and 3.65) and “out-of-cell” ti-
trations were identical (Table 3). This confirms a fast and
quantitative complexation of Ga3+ ions even at pHꢀ1.5,
which was also observed during NMR measurements and
under radiochemical conditions (see below).
In addition, we like to draw attention to another aspect of
the observed kinetics. Although at physiological pH hydrox-
ide anions will compete with the ligand in the [GaACTHNUGRTNEUNG(PrP9)]
cage”) complex, #: intermediates, : free ligand. In the 31P NMR spectra,
*
the small peak at ꢀ42.5 ppm (labeled +) probably corresponds to a dia-
stereomeric complex.
the Supporting Information). Thus, additional NMR signals
could be observed, presumably corresponding to “out-of-
cage” complexes. The chemical shift of the 71Ga NMR signal
at ꢀꢁ30 ppm belongs to some species possessing a rather
symmetric O6 coordination environment that is very proba-
bly formed from phosphinate and carboxylate oxygen atoms
(Figure 4). The transient 31P and 71Ga NMR peaks should
correspond to the “out-of-cage” complexes as interconver-
sion of isomeric “in-cage” octahedral complexes is energeti-
cally demanding and thus does not occur at room tempera-
ture. The 31P NMR peak at ꢀ42.5 ppm was observed in all
complex, this process is very slow compared to the 68Ga
half-life of 68 min. In practice, decay limits the lifetime of
any 68Ga-labeled compound to a couple of hours. For appli-
cations in nuclear medicine, the high “non-equilibrium” sta-
bility constant value of logK[GaL] ꢀ35 that has been mea-
sured in a comparable timeframe is therefore more decisive
than the true long-term thermodynamic stability, since equi-
Chem. Eur. J. 2010, 16, 7174 – 7185
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7179

J. Notni, P. Hermann et al.
final reaction mixtures and can be assigned to a minor dia-
stereomer formed during the process of complex formation.
Moreover, complexation could be observed even under
extremely acidic conditions, that is, in 1, 2, or 5m hydrochlo-
ric acid. In 1m HCl, complete extinction of the [Ga-
pling agents investigated, uronium reagents like HBTU (2-
(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium
hexa-
fluorophosphate) and TBTU (2-(1H-benzotriazole-1-yl)-
1,1,3,3-tetramethyluronium tetrafluoroborate) gave the best
results. Five equivalents of amine and about eight equiva-
lents of coupling agent were found to be sufficient for com-
ACHTUNGTRENNUNG
(H2O)6]3+ signal in the 71Ga NMR spectrum requires about
12 d, whereas in 2m HCl solution an equilibrium was
reached within 30 d (in which approximately 30% of “free”
gallium is present). Finally, in 5m HCl the 71Ga resonance of
the “in-cage” complex could not be detected after a period
of 30 d. However, a 31P NMR resonance of dꢀ40 ppm and
plete derivatization. Subsequently the triACTHGNUTERNNUaG mides 1–4 could
be purified very conveniently by diafiltration through a
membrane with 0.5 kDa cutoff. The compounds, obtained as
the Na+ salts, were easily soluble in water.
Concerning derivatization, the most significant advantage
of PrP9 is that protection of the phosphinate moieties
during the coupling reaction was found to be unnecessary.
Hence, no final deprotection step following conjugation has
to be taken into account, which simplifies conjugation proto-
cols. On the one hand, this could render further protection
of certain targeting vectors unnecessary; on the other hand,
utilization of a vector that does not require any protection
at all offers the possibility of a single-step synthesis of a de-
sired bioconjugate. Moreover, due to the availability of
three equal conjugation sites, the access to multimeric trac-
ers is provided in a very convenient way.
some other signals appeared, indicating that some Ga3+
–
ligand interactions (probably formation of the “out-of-cage”
complex) occurs even in extremely acidic media.
When a sample of [GaACHTUNTRGNEU(GN PrP9)] was dissolved in 6m HClO4
at 258C, no decomplexation was observed over a period of
seven months. A comparable degree of stability has been
observed for [GaACHTUNGTRENNUNG
Hence we conclude that even under harsh conditions, pro-
tons are not able to effectively compete with Ga3+ ion for
the ligandꢆs basic sites (the ring nitrogen atoms), which ren-
ders the complex extremely inert. As observed by potenti-
ometry, the “in-cage” complex, once formed, is protonated
on the phosphinate phosphoryl oxygen atom(s) instead, re-
68Ga labeling: Labeling of the unsubstituted ligand PrP9
with 68Ga was performed by employing various tempera-
tures and pH values, according to an established protocol re-
ported earlier.[41,42] It allows for a very precise adjustment of
pH and temperature and is therefore well suited for the in-
vestigation of labeling properties. Figure 5 shows that at
pH 3, nearly complete (>95%) incorporation of activity is
achieved almost instantaneously at 608C and above. Even at
408C and room temperature, decrease in labeling perfor-
mance is just marginal, requiring three instead of one
minute to reach a >95% plateau. Variation of pH at a con-
stant temperature of 608C revealed that instantaneous label-
ing occurs for pH values from 3–5, whereas for complete ra-
dioactivity incorporation at lower pH, slightly longer reac-
tion times are required. Moreover, for the derivatives 1–4
an almost complete incorporation of activity occurs nearly
as rapidly, as in the case of the unsubstituted compound (see
Supporting Information, Figure S8). Hence we conclude that
functionalization of the side arms does not substantially
affect complex formation; this result is somewhat surprising
but nevertheless quite satisfying.
sulting in [Ga
ed species.
ACHTUNGTRENNUNG
(H4L)]+ (Figure 3) or even multiply protonat-
Conjugation: Clearly, the carboxylate moieties of PrP9 are
predestined to act as conjugation sites for biomolecules.
However, the scope of this study is mainly to deliver proof
of principle, rather than the development of actual PET
tracers. At this stage, we only intended to assess the general
feasibility of such derivatization as well as the radiolabeling
properties of such conjugates. Therefore, triamides of PrP9
with cyclohexylamine and amino acid esters were prepared
as model compounds for conjugates with for example, (oli-
go)peptides (Scheme 2). From some common peptide cou-
The most striking feature of PrP9 and its derivatives is
their ability to incorporate 68Ga3+ at pH values as low as 1,
which has two major advantages. Firstly, at a pH below 3
there is no perturbing formation of the colloidal hydroxide
and all activity is available for radiolabeling in form of free
ACTHNUTRGNEUNG
[68Ga(H2O)6]3+. The second aspect is more a practical one:
The very popular commercially available TiO2-based
68Ge/68Ga generator systems are commonly eluted with 0.1m
HCl, the eluate thus having pH of ꢀ1. None of the common
chelators used for 68Ga radiopharmaceuticals can be suffi-
ciently labeled directly with this eluate. Hence, processing of
the eluate[41] or addition of buffers has been mandatory,
whereby the latter do not play an innocent role in terms of
Scheme 2. Functionalization of PrP9 with amines, leading to conjugates
1–4.
7180
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7174 – 7185

Radiopharmaceuticals
FULL PAPER
chemical incorporation of >95% after 10 min is achieved
by using 14 nmol of the ligand, an amount which is in the
usual range for clinical production of 68Ga tracers.[42,43] This
shows the feasibility of such labeling under conditions
common in clinical practice.
We also performed a direct comparison of PrP9 with
NOTA and DOTA, the most common chelating units in
68Ga chemistry. Here, we employed another commonly used
labeling method.[10] HEPES buffer was used to adjust the
pH of the eluate from a SnO2-based generator (from iThem-
ba LABS, South Africa, eluted with 1m HCl) to a value of
3.3. We chose this pH in order to ensure a fair and realistic
comparison of the ligands, since it has been found to be op-
timal for both DOTA-[43] and NOTA-conjugated peptides.[44]
The results of the labeling at ambient temperature
(ꢀ208C), depicted in Figure 7, show that PrP9 exhibits supe-
Figure 5. Radiochemical yield (RCY; 68Ga incorporation) for labeling of
PrP9 (14 nmol in 5 mL buffered water, 120–160 MBq) upper: pH 3, 20–
1008C; lower: pH 1–5, 608C.
Figure 7. Radiochemical yield (68Ga incorporation) for labeling of PrP9,
NOTA, and DOTA (10 nmol of ligand, 60 MBq in 110 mL HEPES-buf-
fered solution, pH 3.3, 208C). Errors were too small to be displayed at
this scale.
chemical interactions during the labeling process, as well as
with respect to legal and regulatory issues in radiopharma-
ceutical production. We therefore exemplarily performed la-
beling of the benzylglycine conjugate 3 using the neat eluate
from a TiO2-based 68Ga generator (Cyclotron Co., Obninsk,
Russia). Figure 6 illustrates that although radioactivity incor-
poration occurs significantly slower than at pH 3, a radio-
rior performance than DOTA and is also ahead of NOTA,
although the velocity of radioactivity incorporation of the
latter is in a comparable range. It is evident that at lower
pH, the advance of PrP9 will increase because both compet-
itors exhibit drastically reduced labeling yields in more
acidic media.[43,44]
Conclusion
Desirable properties of GaIII chelators for application in
68Ga-based nuclear imaging include high complex stability,
fast and selective complex formation, ability to be conjugat-
ed, long shelf life, and accessibility. Our data show that the
novel ligand system PrP9 introduced herein does fulfill all
of these requirements.
Complex formation kinetics under common chemical as
well as under radiochemical conditions is exceptionally fast,
being superior to the established ligand systems DOTA and
Figure 6. Radiochemical yield (68Ga incorporation) for labeling of
(60 MBq in 400 mL, 608C, pH 1, different molar amounts of 3).
3
Chem. Eur. J. 2010, 16, 7174 – 7185
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7181

J. Notni, P. Hermann et al.
NOTA. This is owed to the assumed complexation mecha-
nism: initially, the Ga3+ ion is coordinated very quickly by
the pendant carboxylates (open-chain-like interaction, lead-
ing to “out-of-cage” complexes). The ion is thereby brought
into direct vicinity of the main chelation site, thus accelerat-
ing transfer to the cage by increasing the effective concen-
tration close to the macrocyclic cavity. Moreover, decom-
plexation under very acidic as well as under neutral and
moderately alkaline conditions is extremely slow. This is of
utmost importance as problems caused by dissociation/trans-
metallation in biosystems can therefore be ruled out, and in
68Ga radioimaging, the risk of uncontrolled distribution of
68Ga3+ activity is eliminated. The fast complex formation
allows for milder labeling conditions (e.g., at ambient tem-
perature), thereby expanding the scope of 68Ga radiolabel-
ing towards targeting vectors with low thermal stability (as
for example, antibodies or their fragments).
All things considered, we believe that PrP9 has truly the
potential to boost the development of 68Ga-based nuclear
medicine and molecular imaging by providing facile and
cost-efficient access to 68Ga radiopharmaceuticals with supe-
rior properties. Further investigations concerning the behav-
ior in biosystems as well as the preparation of conjugates
with targeting vectors are currently under way.
Experimental Section
Materials and methods: All reagents used were of analytical grade. Dry
solvents were used only where indicated and dried according to estab-
lished procedures. The preparation of [9]aneN3 followed essentially a
published procedure,[45] with little alterations as described earlier.[46]
NMR spectra were recorded using a UNITY Inova (400 MHz) or a
VNMRS (300 MHz) spectrometer from Varian. 1H and 13C NMR shifts
are referenced to TMS. 31P NMR shifts are given relative to 85% aq.
H3PO4. 71Ga NMR shifts are referenced to a 1.0m aqueous solution of
Another useful effect of the complexation mechanism is
that formation of colloidal hydroxide is prevented; no for-
mation of precipitates was observed during complexation re-
actions over the entire pH range. Hence, the ligand can be
considered suitable for 68Ga radiolabeling even at pHꢀ5
and above. This expansion of applicable labeling pH range,
together with straightforward labeling at room temperature,
enables its use in radiotracers featuring more sensitive tar-
geting vectors (e.g., antibodies or their fragments). Further-
more, the GaIII complex is quantitatively formed even below
pH 1. This entails the most significant improvement con-
cerning 68Ga radiochemistry, as the pH range for optimal la-
beling is expanded towards quite acidic solutions. For the
first time, 68Ga labeling can be performed using the neat
eluate from TiO2-based 68Ge/68Ga generators (typically 0.1m
HCl, pH 1). Due to possible degradation of sensitive bio-
molecules at such pH, this might finally turn out to be of
lesser practical relevance. However, it means that for PrP9
there is no strict necessity to keep the pH in a very narrow
range from 2.5–4 during the labeling procedures (as re-
quired, for example, for DOTA peptides). This might render
the fully automated syntheses of 68Ga radiotracers more
robust, in which adjustment of labeling pH is often a crucial
issue.
GaACHTNUGRTNEUNG(NO3)3. Elemental analysis was performed using a Heraeus Vario EL
III system. Ultrafiltration/diafiltration was performed using a Millipore
setup (consisting of a 50 mL stirred cell model 8050, CDS10 selector
valve, and RC800 mini-reservoir), in combination with Ultracel cellulose
acetate membranes, filter code YC05, NWML 500Da. Diluted HCl for
elution of 68Ge/68Ga generators was prepared from HCl (suprapure) and
water (ultrapure; both from Merck).
Syntheses: The full synthetic route to of H6PrP9 includes two steps to
obtain the phosphorus precursor; its synthesis has been developed in our
laboratory and published by some of us before.[47] However, as efficacy
and yields were significantly improved, we report the complete protocol
starting from commercially available chemicals, including the mentioned
steps.
Synthesis of H6PrP9: Dry ammonium hypophosphite (0.3 mol, 25 g) and
hexamethyldisilazane (0.48 mol, 77 g, 100 mL) were heated to 1058C
under argon atmosphere with stirring, whereupon gaseous ammonia was
evolved. Caution! The intermediate HPACTHNUTRGNEUNG(OSiMe3)2 is pyrophoric! After
4 h the mixture was cooled to room temperature, dry dichloromethane
(200 mL) was added and the solution was cooled in an ice bath. Then
tert-butyl acrylate (213 mmol, 27.3 g, 31 mL) was added through a syringe
and the mixture was stirred for additional 12 h at room temperature. The
mixture was hydrolyzed by transferring it into ethanol (500 mL) by
means of a capillary and evaporated to dryness. The crude product was
dissolved in chloroform (300 mL) and extracted with two portions
(60 mL each) of 3% aq. HCl. The aqueous phases were combined and
re-extracted with chloroform (3ꢇ50 mL). The combined organic extracts
were dried over anhydrous sodium sulfate and the solvent removed to
afford [2-(tert-butyloxycarbonyl)ethyl]phosphinic acid as a colorless, vis-
cous oil (39.5 g, ca. 203 mmol, 95%), with a purity of 96% according to
31P NMR. The remaining 4% impurity is the P-disubstituted product,
which is inert in the following reactions and thus does not need to be re-
moved. The crude phosphinic acid was dissolved in a mixture of ethanol
(100 mL) and conc. aq. HCl (100 mL) and heated under reflux for 12 h.
Then all volatiles were removed in vacuo to yield (2-carboxyethyl)phos-
phinic acid (34.2 g, ca. 203 mmol, 100%) as a colorless oil which solidifies
upon standing. This compound and 1,4,7-triazacyclononane (45 mmol,
5.8 g) were dissolved in 6m aq. HCl (120 mL) and heated to 708C. Para-
In addition, the synthesis of PrP9 itself is fast, simple, and
scalable; particularly in comparison with other ligand sys-
tems bearing additional carboxylic groups suitable for conju-
gation like DOTAACHTNUTRGENNUG(tBu)3 or NODAGAACHTUNGTERN(NUGN tBu)3. Another
novel and important characteristic is that conjugation by
amide formation needs no protection (and therefore no
final deprotection) as metal coordination sites and function-
al groups for conjugation are different in nature. However,
the most striking advantage of PrP9 is to provide easy
access to multimers. We are not aware of any other system
of such simple structure that could be multiply conjugated
to amine substrates in a single synthetic step. We therefore
hold the view that although monofunctionalization can be
easily achieved as well by performing the coupling reaction
by using a large excess of the ligand, the true designation of
the compound is the preparation of multimeric tracers.
AHCTUNGERTGfNNUN ormaldehyde (0.6 mol, 18 g) was added in small portions with stirring
during a period of 24 h, while progress of the reaction was monitored
with 31P NMR spectroscopy. Then the solvents were distilled off in vacuo
and the remaining HCl was removed by repeatedly adding small portions
of water and evaporating to dryness. The crude product was purified by
chromatography on ion exchange resin (DOWEX 50ꢇ8, H+-form,
column size 25ꢇ6 cm, eluent: water). Impurities were removed with the
first 700 mL of eluate. The next fraction of approx. 1.8 L containing the
pure product was concentrated in vacuo to a volume of 40 mL and meth-
anol (300 mL) was added, whereupon the product slowly crystallized.
7182
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7174 – 7185

Radiopharmaceuticals
FULL PAPER
After 1 h, isopropanol (150 mL) was added and the suspension was
cooled for several hours in order to ensure a complete precipitation. The
product was filtered off, washed with isopropanol and diethyl ether, and
dried in vacuo to yield H6PrP9·2H2O (15.4 g, 55% based on triazacyclo-
Tris(cyclohexylamide)-PrP9 (1): Cyclohexylamine (1 mmol, 100 mg,
115 mL), reaction time: 10 min. Yield: 179 mg (93%); 1H NMR
(400 MHz, D2O): d=1.08–1.27 (m, 15H), 1.52–1.55 (m, 3H), 1.64–1.67
PH
(m, 6H), 1.72–1.80 (m, 12H), 2.30–2.37 (m, 6H; P-CH2), 3.05 (d, J2
=
6.0 Hz, 6H; N-CH2-P), 3.19 (brs, 12H; ring-CH2), 3.45–3.52 ppm (m,
3H); 13C{1H} NMR (75.4 MHz, D2O): d=24.3, 24.9, 26.9 (d, J1PC =92 Hz,
P-C-C), 28.7, 31.9, 48.9, 50.6, 53.6 (d, J1PC =92 Hz, N-C-P), 174.2 ppm (d,
J3PC =17 Hz, C=O); 31P{1H} NMR (161.9 MHz, D2O): d=38.5 ppm; MS
(ESI negative): m/z: 843 [M+Naꢁ2H]+; elemental analysis calcd (%) for
nonane) as
a colorless, fine-crystalline powder. M.p. 218–2198C;
1H NMR (300 MHz, D2O): d=2.07 (dt, J3HH =7.8 Hz, J3PH =13.7 Hz, 6H;
C(O)-CH2), 2.68 (dt, J3HH =7.7 Hz, J2PH =13.2 Hz, 6H; P-CH2), 3.46 (d,
J2PH =5.7 Hz, 6H; N-CH2-P), 3.53 ppm (s, 12H; ring-CH2); 13C{1H} NMR
(75.4 MHz, D2O): d=25.1 (d, J1PC =94 Hz, P-C-C), 26.8 (C(O)-C), 51.5
(ring-C), 54.6 (d, J1PC =89 Hz, N-C-P), 177.3 ppm (d, J3PC =13 Hz, C=O);
31P{1H} NMR (121.4 MHz, D2O): d=40.0 ppm; MS (ESI positive): m/z
(%): 602 (21) [M+Na]+, 626 (100) [M+2NaꢁH]+, 648 (22)
[M+3Naꢁ2H]+; MS (FAB positive): m/z: 580 [M+H]+; elemental analy-
sis calcd (%) for C18H36N3O12P3·2H2O (615.44): C 35.13, H 6.55, N 6.83;
found C 35.04, H 6.60, N 6.67.
Na
found: C 45.20, H 8.35, N 8.74.
Tris{(O-methyl)glycyl}PrP9 (2): Glycine methyl ester hydrochloride
ACHTGNUNERTUN(GN H21)·6H2O (C36H80N6O15P3Na; 952.96): C 45.37, H 8.46, N 8.82;
AHCTUNGTRENNUNG
(1 mmol, 125 mg), reaction time: 45 min. Yield: 138 mg (76%); 1H NMR
(400 MHz, D2O): d=1.70–1.81 (m, 6H; CH2-C(O)N), 2.36–2.45 (m, 6H;
P-CH2), 3.08 (d, J2PH =6.0 Hz, 6H; N-CH2-P), 3.16 (brs, 12H; ring-CH2),
3.68 (s, 9H; CH3), 3.94 ppm (s, 6H; CH2-C(O)O); 13C{1H} NMR
(75.4 MHz, D2O): d=24.7 (d, J1PC =90 Hz, P-C-C), 31.0, 43.9, 53.1, 55.4,
56.0 (d, J1PC =99 Hz, N-C-P), 174.8, 178.6 ppm (d, J3PC =16 Hz, C=O);
31P{1H} NMR (161.9 MHz, D2O): d=38.5 ppm; MS (ESI negative): m/z:
Synthesis of [GaACHTUNGTRENNUNG(H3PrP9)]·nH2O: H6PrP9·2H2O (0.65 mmol, 400 mg)
and GaCl3 (0.65 mmol, 115 mg) were dissolved in water (1 mL). After
brief heating, the pH value of the solution was found to be ꢁ0.4. An
aqueous solution of 5% NH3 was added until a pH of 2.3 was reached
and the complex was allowed to precipitate for 24 h. The solid was fil-
tered off and recrystallized from water (1 mL). The crystals were filtered
off and dried in vacuo, whereupon the material apparently loses a frac-
tion of the co-crystallized water, as the initially large transparent crystals
disintegrate to give 320 mg of a colorless powder. The single crystals of
813 [M+Naꢁ2H]ꢁ; elemental analysis calcd (%) for Na
2ACHTUNGERTN(NUNG H22)·4.8H2O
(C27H59.6N6O19.8P3Na; 901.1): C 35.99, H 6.67, N 9.33; found C 35.98, H
6.35, N 9.19.
TrisACTHNUTRGNE{NUG (O-benzyl)glycyl}PrP9 (3): Glycine benzyl ester hydrochloride
(1 mmol, 202 mg), reaction time: 50 min. Yield: 173 mg (76%); 1H NMR
(400 MHz, D2O): d=1.73–1.80 (m, 6H; CH2-C(O)N), 2.38–2.43 (m, 6H;
P-CH2), 3.05 (d, J2PH =5.6 Hz, 6H; N-CH2-P), 3.14 (brs, 12H; ring-CH2),
3.75 (s, 6H; CH2-C(O)O), 4.85 (s, 6H; CH2-Ph), 7.05–7.12 ppm (m, 15H;
C6H5); 13C{1H} NMR (100.6 MHz, D2O): d=24.1 (d, J1PC =90 Hz, P-C-C),
25.7, 38.6, 47.6, 50.6 (d, J1PC =93 Hz, N-C-P), 64.5, 125.5, 125.8, 126.0,
132.5, 168.4, 173.1 ppm (d, J3PC =16 Hz, C=O); 31P{1H} NMR (161.9 MHz,
D2O): d=38.3 ppm; MS (ESI negative): m/z: 1042 [M+Naꢁ2H]ꢁ; ele-
[Ga
ACHTUNGTRENNUNG
water and those of [GaACHTUNGTRENNUG
into an aqueous solution of the complex. 1H NMR (300 MHz, D2O): d=
2.13–2.21 (m, 6H; C(O)-CH2), 2.69–2.79 (m, 6H; P-CH2), 3.10–3.28 (m,
6H; N-CH2-P), 3.35–3.48 ppm (m, 12H; ring-CH2); 13C{1H} NMR
(101 MHz, D2O): d=21.8 (d, J1PC =100 Hz, P-C-C), 24.2 (C(O)-C), 49.7
(ring-C), 53.3 (d, J3PC =11 Hz, ring-C), 55.4 (d, J1PC =85 Hz, N-C-P),
174.0 ppm (d, J3PC =13 Hz, C=O); 31P{1H} NMR (121.4 MHz, D2O): d=
42.4 ppm; 71Ga NMR (122.0 MHz, D2O): d=135.2 ppm; MS (ESI nega-
tive): m/z: 644/646 [MꢁH]ꢁ; MS (ESI positive): m/z: 668/670 [M+Na]+,
690/692 [M+2NaꢁH]+, 712/714 [M+3Naꢁ2H]+; elemental analysis calcd
(%) for C18H33N3O12P3Ga·1.3H2O (669.53): C 32.29, H 5.36, N 6.28;
found C 32.27, H 5.40, N 6.12.
mental analysis calcd (%) for NaACTHNURGTNE(GNU H23)·5H2O (C45H72N6O20P3Na;
1132.99): C 47.70, H 6.41, N 7.42; found C 47.52, H 6.37, N 7.38.
Tris{(O-tert-butyl)-l-phenylalanyl}PrP9 (4): l-Phenylalanine tert-butyl
ester hydrochloride (1 mmol, 258 mg), reaction time: 50 min. Yield:
223 mg (84%); 1H NMR (400 MHz, D2O): d=1.10 (s, 27H; CH3), 1.73
(brs, 6H; CH2-C(O)N), 2.36 (brs, 6H; P-CH2), 2.77–2.86 (brs, 6H; CH2-
Ph), 3.04 (brs, 6H; N-CH2-P), 3.16 (brs, 12H; ring-CH2), 4.33 (t, J=
4.8 Hz, 3H; CH), 6.97–7.07 ppm (m, 15H; C6H5); 13C{1H} NMR
(100.6 MHz, D2O): d=24.2 (d, J1PC =103 Hz, P-C-C), 24.7, 25.7, 34.6,
47.6, 50.6 (d, J1PC =92 Hz, N-C-P), 52.3, 79.7, 124.0, 125.7, 126.7, 133.9,
169.1, 172.1 ppm (d, J3PC =16 Hz, C=O); 31P{1H} NMR (161.9 MHz,
D2O): d=38.2 ppm; MS (ESI negative): m/z: 1210 [M+Naꢁ2H]ꢁ; ele-
Synthesis of [Fe
ACHTUNGTRENNUNG(H3PrP9)]·nH2O: H6PrP9·2H2O (0.2 mmol, 123 mg) and
[Fe(acac)3] (0.2 mmol, 72 mg) were dissolved in water (0.5 mL) and iso-
ACHTUNGTRENNUNG
propanol (0.1 mL). When heated to reflux, a red solution initially ob-
tained turned into a light yellow one after several minutes. After cooling
to ambient temperature and standing for several hours, yellow crystals
had formed, which were filtered off and recrystallized from water
(0.5 mL). The precipitate was filtered off and dried in vacuo. During
drying, it apparently loses a fraction of the co-crystallized water, as the
initially transparent crystals disintegrate to result in a yellow powder
(88 mg). The single crystals suitable for X-ray diffraction were obtained
mental analysis calcd (%) for Na
CATHUGNTENRN(UNG H24)·6H2O (C57H98N6O21P3Na): C
51.89, H 7.49, N 6.37; found C 51.70, H 7.61, N 6.24.
Crystal structure determination: The diffraction data were collected on a
Nonius Kappa CCD diffractometer (Enraf–Nonius) at 150(1) K with
MoKa radiation (l=0.71073 ꢅ) and analyzed by using the HKL program
package.[48, 49] The structures were solved using direct methods and re-
fined by full-matrix least-squares techniques (SIR92[50] and
SHELXL97[51]). Scattering factors for neutral atoms were included in the
SHELXL97 program. All non-hydrogen atoms were refined anisotropi-
cally. The hydrogen atoms were located in electron density map. Hydro-
gen atoms at carbon atoms were fixed in the theoretical positions. Those
belonging to carboxylate and water oxygen atoms were fixed in original
positions using the riding model with Ueq(H)=1.2Ueq. In the structure of
directly from water as
vapor into the aqueous solution as [FeAHCUTNGTRENNUNG
ACHTUNGTRENNUNG[FeACHTUNGTRENNUNG
tive): m/z: 631 [MꢁH]ꢁ; MS (ESI positive): m/z: 655 [M+Na]+, 677
[M+2NaꢁH]+, 699 [M+3Naꢁ2H]+, 721 [M+4Naꢁ3H]+; elemental anal-
ysis calcd (%) for C18H37N3O14P3Fe (668.26): C 32.35, H 5.58, N 6.29;
found: C 32.20, H 5.63, N 6.09.
Preparation of conjugates—general procedure: H6PrP9·2H2O (0.2 mmol,
123 mg) was dissolved in dry DMSO (3 mL). Then diisopropylethylamine
(DIPEA, 2 mmol, 260 mg, 0.35 mL) was added; in case the coupled
amine was in the hydrochloride form, 3 mmol (390 mg, 0.5 mL) of
DIPEA were used. The amine or amino acid ester hydrochloride
(1 mmol) was added and the mixture was stirred for 5 min. Then TBTU
(1.6 mmol, 0.5 g) was added in small portions within 10 min. The mixture
was left to react for variable time (see individual compound data) and af-
terwards diluted with water (50 mL). For workup, solutions were concen-
trated to 15 mL by ultrafiltration through a membrane with 0.5 kDa
MWCO. Diafiltration with aqueous NaCl (0.05m, 300 mL) removed all
impurities. The solutions were further concentrated to 10 mL and desalt-
ed by diafiltration with pure water (100 mL). Subsequent lyophilization
of the retentate afforded the sodium salts of the products as off-white to
pale yellow, voluminous solids.
[Ga
was best refined disordered in two positions with mutual occupancy of 79
and 21%. CCDC-749080 ([Fe(H3PrP9)]·H2O), -749083 ([Fe-
(H3PrP9)]·6H2O), -749081 ([Ga(H3PrP9)]·2H2O), and -749082 ([Ga-
(H3PrP9)]·6H2O) contain the supplementary crystallographic data for
ACHTUGNTRENN(UGN H3PrP9)]·2H2O, one of the oxygen atoms of one carboxylate moiety
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_re-
quest/cif. For crystallographic parameters see also Supporting Informa-
tion, Table S5.
Potentiometry: The stock/titration solutions used (aq. HCl, ꢀ0.03m;
NMe4OH, ꢀ0.2m) and metal chlorides or nitrates were the same as in
previous studies;[52, 53] the known amount of HCl was added to the stock
Chem. Eur. J. 2010, 16, 7174 – 7185
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7183

J. Notni, P. Hermann et al.
solution of GaCl3 to prevent hydrolysis. Titration conditions: 25.0ꢂ
0.18C; I=0.1m (NMe4Cl); ꢁlog[H+] range of 1.7–11.9 or until precipita-
tion of a metal hydroxide occurred; starting volume 5 mL; ligand concen-
tration ꢀ0.004m; presaturated wet argon as an inert gas. The titration
system consisted of a PHM 240 pH meter, a 2 mL ABU 900 automatic
piston burette and a GK 2401B combined electrode (all Radiometer,
Denmark). At least three parallel titrations were carried out for each
metal-to-ligand molar ratio (1:1 and 2:1); ꢀ40 points per each titration.
In the case of Ga3+ (“equilibrium” titration) and Gd3+ ions, the complex-
ation was too slow for a conventional titration. Thus the “out-of-cell”
method was used: 30 (Ga3+) or 25 (Gd3+) points per titration; ꢁlog[H+]
dissolved in 5m HClO4 The tube was stored at room temperature for
seven months while 31P and 71Ga NMR spectra were recorded regularly.
68Ga labeling—standard labeling procedure:[41] A 10 mL Mallinckrodt
standard glass vial was charged with Millipore water (5 mL) containing
the ligand (14 nmol). Elution of the TiO2-based 68Ge/68Ga generator
(from Cyclotron Co, Obninsk, Russia) was performed with 0.1m HCl
(7 mL). 68Ga activity was concentrated on a small cation-exchange resin
(AG 50-Wꢇ8, 400 mesh, H+-form) and purged with an acetone/HCl-mix-
ture (1 mL). [68Ga]GaCl3 (120–160 MBq, ffi1.2–1.6 pmol) was eluted with
a 2.44% solution of 0.05m HCl in acetone (400 mL) directly into the reac-
tion vessel. The vial was heated to the specified temperature for 20 min.
Samples were withdrawn from the reaction mixture at 1, 3, 10, 15 and
20 min and analyzed by radio-TLC (TLC sheets silica gel 60, mobile
phase: 0.1m sodium citrate in water).
range of 1.5–10.5 (Ga3+/L=1:1), 1.5–3.1 (Ga3+/L=2:1) or 1.8–6.0 (Gd3+
/
L=1:1); starting volume 1 mL; prepared under argon; equilibrium time
at room temperature four weeks (Ga3+ systems) or 1 d (Gd3+ system).
Solutions were kept in tightly closed ground glass tubes below
ꢁlog[H+]<6. For ꢁlog[H+]>6, solutions were flame sealed into am-
poules in order to protect the solutions against atmospheric CO2 during
standing, as alkaline solutions in the closed ground tubes gave irreprodu-
cible results. The solution ꢁlog[H+] in each tube/ampoule for out-of-cell
titration was measured separately with a freshly calibrated electrode (as
given below). The constants determined by this technique showed higher
standard deviations due to less precise measurements and a smaller
number of experimental points. At least two parallel titrations for each
metal-to-ligand ratio (1:1 and 2:1) were performed. The constants (with
standard deviations) were calculated using the OPIUM program.[54–56]
The program minimizes the criterion of the generalized least-squares
method using the following calibration function:
The labeling of 3 with the neat generator eluate was performed by direct-
ly dissolving the specified amount of ligand 3 in the eluate (1.7 mL, typi-
cally containing ꢀ60 MBq of 68Ga, ffi0.6 pmol) and subsequent heating of
the solution to 608C. Probing and analysis were performed as described
above. The stability of 3 under these conditions was checked by heating a
solution of 3 (2 mg) dissolved in HCl (1 mL, 0.1m) to 608C for 20 min.
HPLC analyses (using a 20ꢇ4 mm C18 column, flow rate 1 mLminꢁ1
,
gradient: in 20 min from 30 to 70% MeCN in H2O, both eluents contain-
ing 0.1% trifluoroacetic acid), were performed immediately after dissolv-
ing and after 20 min of heating. The chromatograms showed no signifi-
cant differences and were practically identical to those of the compound
dissolved in pure water.
For labeling in HEPES-buffered solution (comparison of PrP9, NOTA,
and DOTA), a generator with an SnO2 matrix (from iThemba LABS,
South Africa, total eluted 68Ga activity ca. 1500 MBq) was eluted with
1m HCl. A fraction of 1.25 mL containing the highest activity was mixed
with a solution of HEPES (600 mg) in water (Merck ultrapure, 500 mL).
From this solution aliquots of 100 mL (containing about 60 MBq
(ffi0.6 pmol) each) were transferred into eppendorf cups. 10 mL of solu-
tions of the ligand (1 mmolmLꢁ1) were added and mixed well. The pH of
the mixture was 3.3. Samples were withdrawn after 4, 7, 10, 15, and
20 min and analyzed by TLC as described above.
E ¼ E0 þ S log½Hþꢃ þ j1½Hþꢃ þ j2½OHꢁꢃ
in which the additive term E0 contains the standard potentials of the elec-
trodes used and contributions of inert ions to the liquid-junction poten-
tial, S corresponds to the Nernstian slope (the actual value of which
1ACTHNUTRGNEUNG
should be close to the theoretical value) and the j1[H+] and j [OHꢁ]=
j2Kw/[H+] terms are the contributions of the H+ and OHꢁ ions to the
liquid-junction potential. It is clear that j1 and j2 cause deviation from a
linear dependence of E on pH only in strongly acidic and strongly alka-
line solutions. The calibration parameters were determined from titration
of the standard HCl with the standard NMe4OH before each ligand or
ligand–metal titration to give a pair of calibration/titration, which was
used for calculations of the constants. All constants determined are con-
centration constants. The water ion product pKw (13.81) and stability con-
ꢁ
stants of the M2/3+ OH systems included into the calculations were
ꢁ
Acknowledgements
taken from literature.[38,39,57]
NMR measurements: The GaIII–PrP9 complex formation was followed
by 31P and 71Ga NMR spectroscopy at room temperature (258C). Typical
conditions: GaCl3 hydrate (15 mg) was dissolved in distilled water (5 mL)
and the pH was adjusted to 1.3, 1.0 or 0.8 with 0.5m aq. HCl. The ligand
hydrate (10.3 mg) was pre-weighed into a 3 mL vial and quickly dissolved
in this GaCl3 solution (1 mL) to give a solution containing metal ions and
ligand in equimolar amounts. This moment represented zero on the time-
line (t=0). The reaction mixture was transferred as quickly as possible to
a 5 mm NMR tube and placed in the NMR spectrometer. Typically, the
first spectra were obtained at t=2 min. After the end of complexation re-
action, the final pH of the solution was determined again. The 71Ga
NMR spectra were quantified against signal of [Ga(OH)4]ꢁ (5 mm,
pH 13) in the insert tube. Identity of the major isomer in the solution
with that isolated in the solid state was confirmed by addition of the iso-
lated complex into NMR sample after finishing the complexation reac-
tion.
J.N. gratefully acknowledges a postdoc grant from the German Academic
Exchange Service (DAAD). This work was supported by the Grant
Agency of the Czech Republic (grant no. 203/09/1056), the Czech Acade-
my of Sciences (grant no. KAN201110651) and the Ministry of Educa-
tion, Youth and Sport of the Czech Republic (grant no.
MSM0021620857). This work was carried out in the framework of COST
D38 and BM607 Actions and DiMI (grant no. LSHB-2005-512146) proj-
ˇ
ects. We thank Dr. I. Cꢂsarovꢃ for collection of X-ray data.
[1] M. Fani, J. P. Andrꢈ, H. R. Maecke, Contrast Media Mol. Imaging
2008, 3, 67–77.
[2] D. E. Reichert, J. S. Lewis, C. J. Anderson, Coord. Chem. Rev. 1999,
184, 3–66.
[3] M. R. Lewis, D. E. Reichert, R. Laforest, W. H. Margenau, R. E.
Shefer, R. E. Klinkowstein, B. J. Hughey, M. J. Welch, Nucl. Med.
Biol. 2002, 29, 701–706.
[4] A. Anichini, L. Fabbrizzi, P. Paoletti, R. Clay, J. Chem. Soc. Chem.
Commun. 1977, 244–245.
[5] R. D. Hancock, V. J. Thom, J. Am. Chem. Soc. 1982, 104, 291–292.
[6] R. D. Hancock, J. Chem. Educ. 1992, 69, 615–621.
[7] R. D. Hancock, A. E. Martell, Supramol. Chem. 1996, 6, 401–407.
[8] Y. Sun, C. J. Anderson, T. S. Pajeau, D. E. Reichert, R. D. Hancock,
R. J. Motekaitis, A. E. Martell, M. J. Welch, J. Med. Chem. 1996, 39,
458–470.
The dissociation of the GaIII–PrP9 complex in alkaline media was investi-
gated by 71Ga NMR spectroscopy ([GaL]=5 mm, 258C). The complex
was preformed by mixing of the equimolar amount of Ga3+ salt and the
ligand dissolved in water. This solution was evaporated in vacuum and
dissolved in water to give a stock solution of the complex. An increasing
ACTHNUTRGNEUNG
abundance of [Ga(OH)4]ꢁ was quantified against 5 mm [Ga(H2O)6]3+ so-
lution in 0.1m HNO3 in an insert tube. The experiments were carried out
at pH 11.0 (0.2m CAPS buffer) and pH 13 (0.1m NaOH). For decomplex-
ation in acidic media, the complex was preformed in the same way and
7184
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7174 – 7185

Radiopharmaceuticals
FULL PAPER
[9] J. Notni, K. Pohle, J. A. Peters, H. Gçrls, C. Platas-Iglesias, Inorg.
Chem. 2009, 48, 3257–3267.
[38] A. E. Martell, R. M. Smith, Critical Stability Constants, Vol. 1–6,
Springer, Heidelberg, 1974–1989.
[10] W. A. P. Breeman, M. de Jong, E. de Blois, B. F. Bernard, M. Konij-
nenberg, E. P. Krenning, Eur. J. Nucl. Med. Mol. Imaging 2005, 32,
478–485.
[11] W. A. P. Breeman, A. M. Verbruggen, Eur. J. Nucl. Med. Mol. Imag-
ing 2007, 34, 978–981.
[12] A. Heppeler, S. Froidevaux, H. R. Mꢀcke, E. Jermann, M. Bꢈhꢈ, P.
Powell, M. Henning, Chem. Eur. J. 1999, 5, 1974–1981.
[13] P. Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev.
1999, 99, 2293–2352.
[14] The Chemistry of Contrast Agents in Medical Magnetic Resonance
Imaging (Eds.: A. E. Merbach, E. Tꢉth), Wiley, New York, 2001.
[39] NIST standard reference database 46, Version 7.0, Critically Select-
ed Stability Constants of Metal Complexes, 2003.
[40] W. P. Cacheris, S. K. Nickle, A. D. Sherry, Inorg. Chem. 1987, 26,
958–960.
[41] K. P. Zhernosekov, D. V. Filosofov, R. P. Baum, P. Aschoff, H. Bihl,
A. A. Razbash, M. Jahn, M. Jennewein, F. Rçsch, J. Nucl. Med.
2007, 48, 1741–1748.
[42] M. Asti, G. D. Pietri, A. Fraternali, E. Grassi, R. Sghedoni, F. Fioro-
ni, F. Rçsch, A. Versari, D. Salvo, Nucl. Med. Biol. 2008, 35, 721–
724.
[43] G.-J. Meyer, H. Mꢀcke, J. Schuhmacher, W. H. Knapp, M. Hofmann,
Eur. J. Nucl. Med. Mol. Imaging 2004, 31, 1097–1104.
[44] I. Velikyan, H. Mꢀcke, B. Langstrom, Bioconjugate Chem. 2008, 19,
569–573.
ˇ
ˇ
[15] P. Hermann, J. Kotek, V. Kubꢂcek, I. Lukes, Dalton Trans. 2008,
3027–3047.
[16] S. Aime, M. Botta, E. Terreno, Adv. Inorg. Chem. 2005, 57, 173–
237.
[17] M. Van Essen, E. P. Krenning, M. De Jong, R. Valkema, D. J. Kwek-
keboom, Acta Oncol. 2007, 46, 723–734.
[45] J. E. Richman, T. J. Atkins, J. Am. Chem. Soc. 1974, 96, 2268–2270.
[46] J. Notni, H. Gçrls, E. Anders, Eur. J. Inorg. Chem. 2006, 1444–1455.
ˇ
ˇ
ˇ
[47] P. Rezanka, V. Kubꢂcek, P. Hermann, I. Lukes, Synthesis 2008, 1431–
[18] E. T. Clarke, A. E. Martell, Inorg. Chim. Acta 1991, 181, 273–280.
[19] C. Broan, J. P. Cox, A. S. Craig, R. Kataky, D. Parker, A. Harrison,
A. M. Randall, G. Ferguson, J. Chem. Soc. Perkin Trans. 2 1991, 87–
99.
[20] A. S. Craig, D. Parker, H. Adams, N. A. Bailey, J. Chem. Soc. Chem.
Commun. 1989, 1793–1794.
1435.
[48] HKL Denzo and Scalepack Program Package, Z. Otwinowski, W.
Minor, 1997, Nonius BV, Delft, The Netherlands.
[49] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276, 307–326.
[50] A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C.
Burla, G. Polidori, M. Camalli, J. Appl. Crystallogr. 1994, 27, 435.
[51] SHELXL-97, release 97–2, G. M. Sheldrick, 1997, University of
Gçttingen, Gçttingen, http://shelx.uni-ac.gwdg.de/SHELX/.
[21] J. P. Andrꢈ, H. R. Mꢀcke, M. Zehnder, L. Macko, K. G. Akyel,
Chem. Commun. 1998, 1301–1302.
[22] K.-P. Eisenwiener, M. I. M. Prata, I. Buschmann, H.-W. Zhang,
A. C. Santos, S. Wenger, J. C. Reubi, H. R. Mꢀcke, Bioconjugate
Chem. 2002, 13, 530–541.
´
ˇ
[52] P. Tꢃborsky, P. Lubal, J. Havel, J. Kotek, P. Hermann, I. Lukes, Col-
lect. Czech. Chem. Commun. 2005, 70, 1909–1942.
´
[53] M. Fçrsterovꢃ, I. Svobodovꢃ, P. Lubal, P. Tꢃborsky, J. Kotek, P. Her-
[23] P. J. Riss, C. Kroll, V. Nagel, F. Rçsch, Bioorg. Med. Chem. Lett.
2008, 18, 5364–5367.
ˇ
mann, I. Lukes, Dalton Trans. 2007, 535–549.
´
[54] OPIUM, M. Kyvala, 2000, Charles University in Prague, Prague.
[24] T. J. McMurry, M. Brechbiel, C. Wu, O. A. Gansow, Bioconjugate
Chem. 1993, 4, 236–245.
http://www.natur.cuni.cz/kyvala/opium.html.
´
ˇ
[55] CHEMOMETRICS’95, M. Kyvala, I. Lukes, Pardubice, Czech Re-
[25] M. W. Brechbiel, T. J. McMurry, O. A. Gansow, Tetrahedron Lett.
1993, 34, 3691–3694.
[26] I. Lꢃzꢃr, A. D. Sherry, Synthesis 1995, 453–457.
public.
´
ˇ
[56] M. Kyvala, P. Lubal, I. Lukes in IXth Spanish-Italian and Mediterra-
nean Congress on Thermodynamics of Metal Complexes, Girona,
p. 94.
ˇ
[27] K. Bazakas, I. Lukes, J. Chem. Soc. Dalton Trans. 1995, 1133–1137.
[28] J. Huskens, A. Sherry, J. Am. Chem. Soc. 1996, 118, 4396–4404.
[29] C. J. Broan, E. Cole, K. J. Jankowski, D. Parker, K. Pulukkody, B. A.
Boyce, N. R. A. Beeley, K. Millar, A. T. Millican, Synthesis 1992,
63–68.
[57] C. F. Baes, Jr., R. E. Mesmer, The Hydrolysis of Cations, Wiley, New
York, 1976.
[58] H. Stetter, W. Frank, Angew. Chem. 1976, 88, 801; Angew. Chem.
Int. Ed. Engl. 1976, 15, 760.
ˇ
ˇ
[30] I. Lukes, J. Kotek, P. Vojtꢂsek, P. Hermann, Coord. Chem. Rev. 2001,
[59] K. Wieghardt, U. Bossek, P. Chaudhuri, W. Herrmann, B. C. Menke,
J. Weiss, Inorg. Chem. 1982, 21, 4308–4314.
216–217, 287–312.
[31] G. Bandoli, A. Dolmella, F. Tisato, M. Porchia, F. Refosco, Coord.
Chem. Rev. 2009, 253, 56–77.
[60] D. A. Moore, P. E. Fanwick, M. J. Welch, Inorg. Chem. 1990, 29,
672–676.
[32] E. Cole, R. C. B. Copley, J. A. K. Howard, D. Parker, G. Ferguson,
J. F. Gallagher, B. Kaitner, A. Harrison, L. Royle, J. Chem. Soc.
Dalton Trans. 1994, 1619–1629.
[33] D. A. Moore, P. E. Fanwick, M. J. Welch, Inorg. Chem. 1989, 28,
1504–1506.
[34] C.-T. Yang, S. G. Sreerama, W.-Y. Hsieh, S. Liu, Inorg. Chem. 2008,
47, 2719–2727.
[35] M. I. M. Prata, A. C. Santos, C. F. G. C. Geraldes, J. J. P. de Lima, J.
Inorg. Biochem. 2000, 79, 359–363.
[61] E. Arslantas, P. M. Smith-Jones, G. Ritter, R. R. Schmidt, Eur. J.
Org. Chem. 2004, 3979–3984.
[62] S. Liu, E. Wong, V. Karunaratne, S. J. Rettig, C. Orvig, Inorg. Chem.
1993, 32, 1756–1765.
[63] Y. Li, A. E. Martell, R. D. Hancock, J. H. Reibenspies, C. J. Ander-
son, M. J. Welch, Inorg. Chem. 1996, 35, 404–414.
[64] L. G. Luyt, J. A. Katzenellenbogen, Bioconjugate Chem. 2002, 13,
1140–1145.
[36] C. F. G. C. Geraldes, M. C. Alpoim, M. P. M. Marques, A. D. Sherry,
M. Singh, Inorg. Chem. 1985, 24, 3876–3881.
[37] G. Anderegg, F. Arnaud-Neu, R. Delgado, J. Felcman, K. Popov,
Pure Appl. Chem. 2005, 77, 1445–1495.
Received: November 30, 2009
Revised: February 4, 2010
Published online: May 12, 2010
Chem. Eur. J. 2010, 16, 7174 – 7185
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7185