A New Family of Insulin-Mimetic Vanadium Complexes Derived from 5-Carboalkoxypicolinates

Article abstract of DOI:10.1002/chem.200305019

The reaction of 5-carboalkoxypicolinic acid (5ROpicH, R= Me, Et, iPr, sBu; 1a-d) with vanadyl sulfate yielded the complexes [VO(H ₂O)(5ROpic)₂], 2a-d, with H₂O and one of the picolinato ligands in the equatorial positions, and the second picolinate occupying equatorial (N) and axial (O) positions. Reaction of 1a with [NH ₄][VO₃] yielded [NH₄][VO ₂(5MeOpic)₂], [NH₄]-3, in which the N functions of the picolinates are trans to the doubly bonded, cis-positioned oxo groups. Complexes 1a · H₂O, 1b, 1c, 2a · 3.5H ₂O and [NH₄]-3 · 4H₂O have been structurally characterised. A detailed pH-potentiometric solution speciation analysis of the system VO²⁺-1a revealed a dominance of VO(5OMepic)₂ between pH 2 and 6, with the same coordination pattern, evidenced by EPR spectroscopy, as in the crystalline solid state. In ternary systems containing physiological concentrations of the low molecular mass biogenic binders (B) lactate, oxalate, citrate or phosphate, ternary species of general composition VO(5MeOpic)B dominate at physiological pH, with citrate being the most effective competitor for picolinate. All of the complexes trigger glucose uptake and degradation by simian virus modified mice fibroblasts at non-toxic concentrations (< 100 μM), with 2 a, [VO ₂(pic)₂]^- and [VO₂(dipic)] ^- being at least as effective as insulin. Vanadium uptake by the cells is most effective in the case of 2a. 2a also effectively inhibits free fatty acid release by rat adipocytes treated with epinephrine, thus mimicking the inhibition of lipolysis by insulin.

Full text of DOI:10.1002/chem.200305019

FULL PAPER
A New Family of Insulin-Mimetic Vanadium Complexes Derived from
5-Carboalkoxypicolinates
[
a]
[b]
[c]
[c]
[d]
Jessica G‰tjens, Beate Meier, Tam a¬ s Kiss, Eszter M. Nagy, Peter Bugly o¬ ,
[
e]
[e]
[a]
Hiromu Sakurai, Kenji Kawabe, and Dieter Rehder*
Abstract: The reaction of 5-carboalkox-
ypicolinic acid (5ROpicH, R ˆ Me, Et,
iPr, sBu; 1a ± d) with vanadyl sulfate
pH-potentiometric solution speciation
VO(5MeOpic)B dominate at physiolog-
ical pH, with citrate being the most
effective competitor for picolinate. All
of the complexes trigger glucose uptake
and degradation by simian virus modi-
fied mice fibroblasts at non-toxic con-
centrations (<100 mm), with 2a,
2‡
analysis of the system VO -1a revealed
a dominance of VO(5OMepic)2 be-
tween pH 2 and 6, with the same coor-
dination pattern, evidenced by EPR
spectroscopy, as in the crystalline solid
state. In ternary systems containing
physiological concentrations of the low
molecular mass biogenic binders (B)
lactate, oxalate, citrate or phosphate,
ternary species of general composition
yielded the complexes [VO(H O)(5R-
2
Opic) ], 2a ± d, with H O and one of the
2
2
picolinato ligands in the equatorial po-
sitions, and the second picolinate occu-
pying equatorial (N) and axial (O)
positions. Reaction of 1a with
À
À
[VO (pic) ] and [VO (dipic)] being at
2
2
2
least as effective as insulin. Vanadium
uptake by the cells is most effective in
the case of 2a. 2a also effectively
inhibits free fatty acid release by rat
adipocytes treated with epinephrine,
thus mimicking the inhibition of lipolysis
by insulin.
[
NH ][VO ] yielded [NH ][VO (5-
4 3 4 2
MeOpic) ], [NH ]-3, in which the N
2
4
functions of the picolinates are trans to
the doubly bonded, cis-positioned oxo
Keywords: bioinorganic chemistry ¥
groups. Complexes 1a ¥ H O, 1b, 1c, 2a ¥
2
diabetes
¥ picolinates ¥ solution
3
.5H O and [NH ]-3 ¥ 4H O have been
2 4 2
speciation ¥ vanadium
structurally characterised. A detailed
[
6]
Introduction
(and, more recently, also ‡iii ) that exhibit insulin-mimetic
activities and, concomitantly, low toxicity. Vanadate itself is
toxic due to its inhibitory effect towards various phosphate-
The use of sodium metavanadate in the treatment of human
diabetes mellitus had already been reported as early as 1899,[1]
but interest in pharmaceutically available insulin substitutes
[7]
metabolising enzymes. The main advantage of vanadium
compounds, as compared to insulin, lies in the possibility of
oral administration. Furthermore, vanadium compounds are
active in type 1 diabetes (which is due to insufficient or
lacking insulin supply) and type 2 diabetes (non-insulin-
dependent diabetes mellitus, due to insulin resistance).
The insulin-like activities exhibited by vanadium com-
pounds encompass the ability to lower elevated blood glucose
levels in vivo. Vanadium stimulates glucose uptake and
subsequent oxidation in glucose-metabolising cells as well as
glycogen synthesis in vitro. In addition, it promotes the
inhibition of glycolysis and glyconeogenesis, and inhibits
lipolysis (activates lipogenesis).
containing vanadium has revived only during the last
2±4]
2
5 years.[
Efforts have been made to design coordination
compounds of vanadium in the oxidation states ‡iv and ‡v
[5]
[
a] Prof. Dr. D. Rehder, Dipl.-Chem. J. G‰tjens
Institute of Inorganic and Applied Chemistry
University of Hamburg, 20146 Hamburg (Germany)
[
b] Prof. B. Meier
BCM-Research Institute
5307 Entrischenbrunn (Germany)
8
[
c] Prof. T. Kiss, E. M. Nagy
Department of Inorganic and Analytical Chemistry
University of Szeged, 6701 Szeged (Hungary)
To be therapeutically effective, a compound should exhibit
a balanced lipophilicity and hydrophilicity, which can be
achieved by an appropriate design of the periphery of the
coordination compound, thereby providing facile uptake by
and transport in the bloodstream. The complexes should
further be stable with respect to ligand exchange in order to
survive the acidic conditions in the stomach and the transport
[
d] Dr. P. Buglyo¬
Department of Inorganic and Analytical Chemistry
University of Debrecen, 4010 Debrecen (Hungary)
[
e] Prof. Dr. H. Sakurai, K. Kawabe
Department of Analytical and Bioinorganic Chemistry
Kyoto Pharmaceutical University, Kyoto, 607-8414 (Japan)
4924
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200305019
Chem. Eur. J. 2003, 9, 4924 ±4935

4
924±4935
by body fluids. To facilitate efficient transport across hydro-
phobic membranes into the target cell, the ligand sphere
should also contain a site for recognition by a membrane
receptor. Finally, non-toxic metabolites are advantageous.
This latter requirement is fulfilled with picolinic acid (pyr-
idine-2-carboxylic acid), which is an intermediate in the
tryptophane degradation path way.
Promising results with regard to the aspects mentioned
above have thus been obtained with vanadium complexes
bearing ligands derived from picolinic acid, picH. These
ligands form vanadium compounds with the NO, OO or ONO
coordination modes; for an early review by J. Selbin reference
Results and Discussion
Synthesis, characterisation and structure description:
Scheme 1 provides an overview of the ligands and complexes
introduced here. For the syntheses of the 5-carboalkoxypico-
linic acids, 5ROpicH (1a ± d), the pyridine-2,5-dicarboxylic
acids were converted to the dialkylesters according to known
procedures.[ Reaction with copper(ii) nitrate yielded blue
copper complexes, from which the monoesters were liberated
with hydrogen sulphide.[ In the case of 1a, 1b and 1c single
crystals were grown out of aqueous acetone. The acids
14]
15]
≈
crystallise in the triclinic space group P1 (1a ¥ H O, 1b) or in
2
[
8]. Complexes with vanadium in oxidation state ‡iv and
the monoclinic space group P2(1)/c (1c). The molecular
structures and sections of the crystal structures showing the
H-bonding network are depicted in Figure 1. Bond lengths
picolinate as ligand, [VO(pic) ], were initially tested for their
insulin-like activity in 1995 by Sakurai and co-workers, who
demonstrated the ability of these complexes to normalise both
2
and angles (Table 1) are in the expected range. In 1a ¥ H O,
2
[9]
serum glucose and free fatty acid (FFA) levels in rats.
adjacent molecules of 1a are linked together through two
water molecules by hydrogen bridges; the water molecules act
as donors to the pyridine-N and as acceptors for the
Starting from these results, various VIV and VV compounds
with picH and derivatives of picH were the subject of research
activities in terms of their insulin-mimetic (normoglycemic)
carboxylic-OH, giving rise to chains consisting of H O-
2
[10]
activity. Examples are complexes with picolineamides, and
picolinates with the substituents 5-I, 6-Me, 6-COO (dip-
ic),
the oxidation state of vanadium does not seem to play a
crucial role in activity.[
bridged dimers of 1a. These chains are interlinked by
hydrogen bonds between water and the carboxylic-O. In 1b,
trimeric subunits are formed by OH ¥¥¥ N links, while in 1c, the
molecules are infinitely linked through OH ¥¥¥ N hydrogen
bonds.
[11]
À
[
10, 12, 13]
[14, 15]
[16, 17]
6-Et
and 4-OH-dipic.
It is noteworthy that
18]
The complex
anion
À
[VO (dipic)(H O)] proved promising in the oral treatment
For the syntheses of the vanadium(iv) complexes 2a ± d, an
aqueous solution of vanadyl sulfate was mixed with a solution
of 5ROpicH and sodium acetate in water/THF and heated to
508C for 2 h. The complexes precipitated in the form of green,
microcrystalline powders. Once generated, the complexes are
stable in air. Elemental analyses confirmed the expected
coordination of two molecules of the ligand to the metal
centre, with the sixth position occupied by a molecule of
water. In the case of [VO(H O)(5MeOpic) ] (2a), green
2
2
[19, 20]
of diabetic cats.
The peroxo complex K [VO(O ) pic] has
2 2 2
been shown to decrease blood glucose levels in streptozoto-
cin-induced diabetic rats.[ Apart from the various picolinato
vanadium compounds, there also exist several examples for
picolinato complexes with metal centres other than vanadium,
such as zinc, cobalt and chromium, which have also been
shown to be insulin-mimetic.[
Here, we present a novel variant of the picolinato scheme,
with regard to vanadium complexes of 5-carboalkoxypicolinic
acid, 5ROpicH (Scheme 1), with a site for modification, with
respect to the substituent R in the 5-position, which allows a
fine tuning of the properties of the respective vanadium
complexes in the periphery of the coordination sphere.
21]
3, 21]
2
2
single crystals of 2a ¥ 3.5H O were obtained from a hot
2
aqueous solution by slow cooling. The yellow bis(oxo)vana-
dium(v) complex [NH ][VO (5MeOpic) ], [NH ]-3, was ob-
4
2
2
4
tained by mixing a slurry of [NH ][VO ] in water with a
4
3
solution of the 5MeOpicH in acetone. Single crystals of
NH ]-3 ¥ 4H O suitable for an
[
4
2
X-ray diffraction analysis were
grown from the mother liquid.
The molecular structure of
[
VO(H O)(5MeOpic) ] (2a) is
2
2
displayed in Figure 2, with
bonding parameters in Table 2
and, for the ligand, in Table 1.
The asymmetric unit contains
two independent molecules and
seven waters of crystallisation.
The basic formula of the crys-
talline material thus is 2a ¥
3
.5H O and crystallises in the
2
≈
triclinic space group P1, while
the bulk material, as evidenced
by the elemental analysis and
thermogravimetry (vide infra),
is free of water of crystallisa-
tion. In one of the independent
Scheme 1.
Chem. Eur. J. 2003, 9, 4924 ±4935
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4925

D. Rehder et al.
FULL PAPER
2
Figure 1. XSHELL plots (50% probability level) of 1a ¥ H O, 1b and 1c.
molecules, the methoxy group is disordered over two posi-
tions (70:30%). Four of the waters of crystallisation are
disordered as well. See also Experimental Section for disorder
information. The ligands coordinate in the usual bidentate
manner, forming, together with an aqua ligand (O2) in the
equatorial and the doubly bonded oxygen (O1) in the apical
position, a complex of distorted octahedral geometry. The
second axial position is occupied by the carboxylate oxygen
O7 of one of the picolinates. The corresponding bond length
d(V1ÀO7), 2.122(3) ä, is rather long as a consequence of the
The ligand bonding parameters (Table 1) are the same within
the limits of error as in the free acids, with the exception of the
bond length carboxylate-OH (C1ÀO1 in 1a, which is some-
what longer than for the deprotonated, coordinated carbox-
ylate-O (C1ÀO3 and C9ÀO7) in 2a.
[NH ][VO (5MeOpic) ] ¥ 4H O ([NH ]-3 ¥ 4H O) (Figure 3
4
2
2
2
4
2
and Table 2) crystallises in the monoclinic space group C2/c.
Two of the water molecules are disordered, one to the extent
where it is displaced along the C2 axis across the space
between the facing VO moieties of two neighbouring anions;
2
trans influence of O1. The angle O1-V-O7, 165.1(1)8, signifi-
cantly deviates from linearity, possibly as a result of the steric
hindrance imparted by one of the carbomethoxy groups.
Vanadium is displaced from the equatorial plane spanned by
N1, N2, O2 and O3 towards the apical O1 by 0.295 ä. Water
of crystallisation is hydrogen-bonded to the aqua ligand and
the coordinated carboxylate-oxygen atoms O3 and O7, and
further forms a weak bond to the methoxy group O5; Table 2.
Figure 3, right. The molecular structure is shown in Figure 3,
left. The oxo groups (O5, O5A) are cis standing, bisected by
the C axis which passes through the vanadium centre and the
2
octahedral edge formed by N1 and N1A. In contrast to 2a,
where a carboxylate-O is trans to the oxo group, the pyridine-
N are trans to the oxo groups in 3, giving rise to rather long
d(VÀN1) ˆ 2.310(2) ä. The molecular axis, defined by O5-V-
N1, is again tilted: the angle is 163.42(7)8. Vanadium is above
4926
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4924 ±4935

Insulin-Mimetic Vanadium Complexes
4924±4935
Table 1. Selected bond lengths [ä] and angles [8] for 1a ¥ H
2
O, 1b and 1c, and the ligand 1a(1 À ) in the complexes 2a and 3.
À
À
1
a ¥ H
2
O
1a in 2a ¥ 3.5H
2
O
1a in [NH
4
]3 ¥ 4H
2
O
1b
1c
C1ÀC2
1.5058(18)
1.3110(16)
1.2148(16)
1.4984(19)
1.2009(17)
1.3323(16)
1.4543(17)
125.37(12)
124.56(12)
114.91(11)
1.511/1.502(5)
1.504(3) C1ÀC2
1.284(2) C1ÀO1
1.226(2) C1ÀO2
1.490(3) C5ÀC7
1.195(3) C7ÀO3
1.318(3) C7ÀO4
1.464(3) C8ÀO4
1.5093(17)
1.5064(14)
C1ÀC2/C9ÀC10
C1ÀO1
1.304/1.279(5)
1.3066(15)
1.2117(15)
1.4928(17)
1.2111(15)
1.3292(15)
1.4622(15)
125.93(11)
124.45(11)
114.27(10)
1.3188(13)
1.2068(14)
1.4920(14)
1.2079(13)
1.3336(13)
1.4688(12)
125.84(10)
124.91(10)
116.19(8)
C1ÀO3/C9ÀO7
C1ÀO2
1.232/1.232(4)
C1ÀO4/C9ÀO8
C5ÀC7
1.498/1.481(5)
C5ÀC7/C13ÀC15
C7ÀO3
1.208/1.209(4)
C7ÀO5/C15ÀO9
C7ÀO4
1.327/1.324(5)
C7ÀO6/C15ÀO10
C8ÀO4
1.456/1.454(5)
C8ÀO6/C16ÀO10
O1-C1-O2
O3-C7-O4
C7-O4-C8
126.6/124.9(4)
124.29(19) O1-C1-O2
123.8(2) O3-C7-O4
116.2(18) C7-O4-C8
O3-C1-O4/O7-C9-O8
124.7/124.9(4)
O5-C7-O6/O9-C15-O10
115.3/115.7(3)
C7-O6-C8/C15-O10-C16
O1 ¥¥¥ O01
O01 ¥¥¥ O2
O01 ¥¥¥ N1
O1 ¥¥¥ N1
2.52
2.77
2.83
2.70
2.72
Table 2. Selected bond lengths [ä] and angles [8] for 2a ¥ 3.5H
2
O and
4 2
]-3 ¥ 4H
O.[
a]
[
NH
2
a ¥ 3.5H
2
O
4 2
[NH ]-3 ¥ 4H O
V1ÀO1
V1ÀO2
V1ÀO3
V1ÀO7
V1ÀN1
V1ÀN2
1.597(3)
2.016(3)
1.989(3)
2.122(3)
2.118(3)
2.133(3)
165.04(12)
161.37(13)
163.31(12)
78.71(11)
74.53(11)
2.600(5)
2.688(5)
2.832(5)
2.886(5)
3.065(6)
V1ÀO5
V1ÀO1
V1ÀN1
1.6287(14)
1.9784(14)
2.3098(16)
O1-V1-O7
O2-V1-N1
O3-V1-N2
O3-V1-N1
O7-V1-N2
O2 ¥¥¥ O21
O2 ¥¥¥ O22
O21 ¥¥¥ O7
O22 ¥¥¥ O3
O23 ¥¥¥ O5
O5-V1-N1
O1-V1-O1A
O5-V1-O5A
O1-V1-N1
163.42(7)
150.13(8)
105.68(11)
73.83(5)
O10 ¥¥¥ O3
O10 ¥¥¥ O5
N2 ¥¥¥ O2
2.803(3)
2.896(2)
2.927(2)
2.843(2)
Figure 2. XSHELL plot (50% probability level) of 2a. Only one of the two
independent molecules is shown.
N2 ¥¥¥ O10
[
a] O21, O22 and O10 correspond to water of crystallisation.
the plane spanned by O1, O1A, N1A and O5A by 0.296 ä.
The d(VÀO5), 1.629(1) ä, are relatively long, possibly as a
ligand parameters compare to those of the free picolinic acid,
consequence of the hydrogen bonding to water of crystalli-
except again for a shortening of d(C1ÀO1) in 3 as compared to
sation, O10, and reflected in the rather low value n(VˆO); see
1a, Table 1.
Table 3. O10 is further hydrogen-bonded to the ammonium
cation, which in turn forms a hydrogen bond to the non-
coordinated carboxylate-O2. In addition, O3 of the ester
group forms hydrogen bonds with interstitial water. The
Selected IR data are collated in Table 3. The CˆO stretch is
somewhat strengthened in the complexes, reflected by a shift
À1
to higher wavenumbers by approximately 15 cm in the
complexes with respect to the free ligands. The difference
Table 3. Selected IR[a]data.
1
a
1b
1c
1d
2a
2b
2c
2d
3
n(CˆO)
1719, 1698
1717, 1697sh
1717
1714
1733
1654
1397
971
1737
1651
1402
978
1730
1652
1397
967
1729
1650
1403
967
1732
1658
1396
903br
n(CO
n(CO
2
)
)
as
2
sym
n(VˆO)
a] KBr, in cm
Chem. Eur. J. 2003, 9, 4924 ±4935
À1
[
.
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4927

D. Rehder et al.
FULL PAPER
4 2
Figure 3. Molecular structure (left; 50% probability level) and section of the crystal structure (right) of [NH ]-3 ¥ 4H O.
between the antisymmetric and symmetric carboxylate bands
corresponds to what one expects for the monodentate
coordination of carboxylates. The VˆO stretch is normal for
Table 4. EPR parameters and suggested binding sets for the complexes formed
2
‡
2‡
in the binary system VO -5MeOpicH (AH) and the ternary systems VO -AH-
BH , BH
n
n
ˆ oxalic/lactic/citric/phosphoric acid.
the VO2 complexes 2a ± d; it is rather low for the VO
‡
‡
Species[a]
Ligand sets[b]
g
z
A
z
2
À4
À1
(
10 cm )
complex 3, which may be due to the above-mentioned
intimate hydrogen bonding to water of crystallisation.
‡
À
[VOA]
I
1.939 179
1.947 168
1.947 171
1.948 167
(N, CO
(N, CO
(N, CO
(N, CO
2
2
2
2
)
À
À
À
À
cis-[VOA
VOAHÀ1
VOAHÀ2
2
] II
); (N, CO
); H
); OH ; OH
2
ax); H
2
O
Thermogravimetry shows the loss of one H O molecule
2
À
À
[
[
]
III
2
O; OH
(
4.1%) at 1308C, which is typical for vanadyl complexes
2
À
À
À
]
IV
[25]
containing an equatorially coordinated aqua ligand. The
first ligand is degraded between 240 and 3808C, the second
around 5008C. The overall weight loss due to ligand degra-
dation amounts to 69.7%, that is, part of the picolinate ligands
oxalic acid
cis-[VOAB]
lactic acid
À
À
À
À
1.943 170
1.946 166
(N, CO2 ); (CO2 , CO2 ); H O
ax
2
À
À
cis-[VOAB]
(N, CO
2
); (CO
2
, OHax); H
2
O
citric acid
(
corresponding to 4H, 2O, and 1N) remains on vanadium,
2À
À
À
À
À
À
ax
[
VOAB]
1.945 169
1.948 159
(N, CO
(N, CO2
2
); (CO
); (CO2
2
, OH
, O ,
,
CO
CO2
2
)
which is recovered in the form of [NH ][VO ]. The third
[VOABH-1]3À
À
À
ax)
4
3
oxygen necessary to generate metavanadate apparently stems
from external oxygen.
phosphoric acid
(2Àx)À, x ˆ 2,1,0 1.940 177
(N, CO
À
); H
(3Àx)À
4
[
VOABH
x
]
2
x
PO
2
; H O
[
a] See also Table 5 for the species I ±IV; cis refers to the position of the aqua
ligand (cis with respect to the doubly bonded O). [b] Equatorial, if not indicated
otherwise.
Solution speciation studies: To assess the stability of the
complexes under various pH conditions and in the presence of
competitive low molecular mass ligands present in the blood
serum (oxalate, lactate, citrate, phosphate), we have carried
out speciation studies for the systems containing 5-methyl-
carboxy-picolinic acid (1a; abbreviated AH in the context of
the speciation studies) by pH potentiometry. Binding modes
of the complexes in solution were determined by EPR
spectroscopy.[ The results are represented in Table 4, and
Table 5. EPR parameters of the binary system VO2 -5MeOpicH (AH).
‡
pH Species[a] (ratio[b]
)
g value
A value (10 cm )
À4
À1
g
0
g
z
g
xy
A
0
A
z
A
xy
1
4
.85 I (90%)
II (10%)
.05 I (60%)
II (40%)
1.968 1.939 1.982 96.7
1.968 1.947 1.978 96.7
1.970 1.939 1.986 91.9
1.970 1.946 1.983 91.9
1.972 1.947 1.985 93.9
1.971 1.948 1.983 90.4
1.972 1.948 1.985 89.5
179.8
168.0
179.1
168.4
171.0
167.6
166.7
55.1
61.0
48.2
53.6
55.4
51.8
50.9
26]
2‡
in Table 5 in more detail for the binary VO -AH system.
According to these findings, the species [VOA ], which
6.00 III (100%)
7
8
2
.42 IV (100%)
.80 IV (100%)
dominate over a wide pH range (vide infra), attain the same
ligand distribution as is found for the solid state structure; see
2
a in Figure 2 and Scheme 1.
[a] See Table 4. [b] Roughly estimated from peak height.
The protonation constants (logK) for 1a, and the stability
constants for the complexes (logb) are listed in Table 6,
together with those of picolinate and 6-methylpicolinate[
for comparison (see Experimental Section for notation). The
presence of the carbomethoxy group in 1a significantly
decreases the basicity of both protonation sites, the pyri-
dine-N and the carboxylate, as a consequence of the electron-
27, 28]
withdrawing effect of the ester group. Compound 1a, similar
to picH and 6MepicH, forms mono- and bis(ligand) com-
2
‡
plexes of different protonation states with VO (Table 6 and
Figure 4). The stability constants for the complexes formed
with 1a(1 À ) are about the same as those for 6Mepic(1 À ).
4928
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4924 ±4935

Insulin-Mimetic Vanadium Complexes
4924±4935
[VOH A] . On further increasing the pH, deprotonation
À
Table 6. Protonation constants (logK), and stability constants for the com-
À1
2
‡
plexes (logb) formed between VO and A(H) at 258C and I ˆ 0.2m (3s errors
2À
proceeds by formation of [VOH A] (species IV in Table 4),
À2
in parentheses).
until finally ligand-free and EPR-silent vanadylhydroxides of
higher nuclearity are generated.
5
MeOpicH, 1a picH[b] 6MepicH[b]
logK(HA)[a]
logb[VOA]
logb[VOHÀ1A]
3.35(1)
5.16(4)
0.38(4)
3.75(4)
À 6.8(2)
9.52(4)
4.36
5.19
6.66
±±
6.15
±
12.11
5.45
1.21
5.82
5.13
Due to their high affinity for hard (oxophilic) metal ions,
citrate, oxalate, lactate and phosphate are competitive binders
‡
2‡
for VO in blood serum. The serum concentrations of these
logb[(VO)
2
H
À2
A
2
]
3.25
À 6.56
9.28
4.15
0.98
À
low molecular mass biogenic ligands amount to 0.01 mm for
oxalate, 0.1 mm for citrate, 1.5 mm for lactate and 1.1 mm for
logb[VOHÀ2A]
logb[VOA
logK(VOA
2
]
phosphate.[ The stability data for the ternary VO -A -B
29]
2‡
À
nÀ
2
)
logK(VOA/VOA
2
)
0.80
1.81
2.82
À
nÀ
systems (where A is 5MeOpic and B the competitive
ligand anion) are summarised in Table 7; for EPR data and
ligand sets derived thereof see Table 4. Speciation diagrams
are presented in Figure 5.
logK for VO2 ‡ HA ˆ [VOA] ‡ H
‡
‡
‡
1.47 À 0.69
logK for VO2 ‡ 2HA ˆ [VOA
‡
‡
1.73 ±2.36
2
] ‡ 2H
[
a] Protonation site for the pyridine-N. The logK for the protonation of the
carboxylate is ꢀ1 for picH and 6MepicH, and <1 for 5MeOpicH. [b] From
In the ternary system containing oxalate, the potentiomet-
ref. [27].
ric titration curves could be fitted well by assuming the
À
formation of two new complexes, [VOAB]
and
However, the equilibrium constants for the proton displace-
ment reactions shown in the two last rows in Table 6 clearly
2
‡
Table 7. Stability constants of mixed ligand VO complexes at 258C and I ˆ
2‡
indicate that 1a(1 À ) coordinates to VO significantly more
readily than 6Mepic(1 À ). This should be mainly due to the
2‡
0
.2m; VO :HA:H
n
B ˆ 1:2:2; c(V) ˆ 1 mm (3s errors in parentheses).
À
À
3À
3À
Complex
Oxalate (B2
)
Lactate (B ) Citrate (B
)
Phosphate (B )
steric hindrance of the methyl group adjacent to the pyridine-
N in the latter.[
27, 28]
Figure 4 shows that the mono- and
[VOABH2]
±2 ±6 .3(1)
±1 ±5 .43(3)
11.33(1)
3.89(5)
0.0092
[
[
[
VOABH]
VOAB]
VOABHÀ1
22.70(7)
bis(ligand) species form simultaneously, with [VOA ] domi-
2
8.16(4)
4.08(2)
0.0087
11.84(3)
7.38(2)
0.0102
16.51(8)
10.13(7)
0.0134
248
nating the pH range 2 ±6. In the pH range 5 ±7, deprotonation
]
(
of aqua ligands) is accompanied by an about 50% intensity fitting[a]
loss of the EPR signal. This can be interpreted by the no. of points 296
314
316
formation of the EPR silent, hydroxo-bridged dinuclear
[
a] The average difference between the calculated and experimental titration
2À
[
(VO) H A ]
2 À2 2
species along with the mononuclear curves expressed in mL of the titrant.
Figure 4. Speciation of the complexes formed in the VO2‡-AH system (AH ˆ 5MeOpicH, 1a) at a 1:4 metal-to-ligand ratio and c(V) ˆ 1 mm.
Chem. Eur. J. 2003, 9, 4924 ±4935
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4929

D. Rehder et al.
FULL PAPER
2
‡
À
nÀ
Figure 5. Speciation of the complexes formed in the VO -AH-BH
n
systems [A ˆ 5MeOpic, 1a, B (from top left to bottom right) ˆ oxalate, lactate,
citrate, phosphate] at a 1:2:2 (VO :A :BnÀ) ratio and c(V) ˆ 1 mm.
2
‡
À
2
À
À
À
2
[
VOH AB] , dominating the pH range 2 ±8. The formation
species is thus {Naryl, CO , CO , H O}. The formation of
2 2
À1
‡
of the mixed complexes according to the reactions VOA ‡
ternary systems is also favoured with citrate. In this system,
the curves can be fitted with the species [VOABH] ,
2À
À
À
À
À
B
(
! [VOAB] (logK ˆ 6.17) and VOB ‡ A ! [VOAB]
2À
3À
logK ˆ 5.44) is clearly favoured over the formation of the
[VOAB] and [VOABH ] . Based on the similarity of the
À1
2À
[30]
complexes [VOB ] (logK ˆ 4.79 ) and [VOA ] (logK ˆ
EPR parameters in this system with those of the pic and
2
2
[27, 28]
À
4
(
.36). The EPR data indicate the presence of a single species
Table 4) over the pH range 2 ±7. The apparent contradiction
with the speciation diagram (Figure 5), which indicates the
6Mepic systems,
an equatorial ligand set {N , CO ,
aryl
2
À
2À
CO , OH} is suggested for [VOAB] below pH ꢀ 4, and
2
À
À
À
3À
{Naryl, CO , CO , O } for [VOABH ] above pH ꢀ 4, that
2
2
À1
2À
simultaneous presence of about 15% of the species [VOB ] ,
can be explained with the very similar contributions of the
CO2 and Naryl donor atom sets to the parallel (z) component
of the hyperfine coupling constant yielding basically the same
A values for both the ternary species [VOAB] and the binary
bis(oxalato) complex [VOB ] . The recorded set of EPR
parameters is in good agreement with those for the ternary
is deprotonation and coordination of the alcoholic group. The
significantly lower value of pK(VOAB) for 5MeOpic (4.42)
compared to pic (5.76) or 6Mepic (5.66) indicates the more
2
À
2‡
favoured formation of this complex in the VO -5MeOpicH
À
2À
system. The decrease of A on going from [VOAB] (169) to
z
z
2À
3À
À4
À1
[VOABH-1] (159 Â 10 cm ) confirms the presence of a
2
stronger donor set for the latter. Ternary complexes
À4
À1
complexes of pic (g ˆ 1.941; A ˆ 167 10 cm ) or 6Mepic
[VOAHB] and [VOAH B] dominate with phosphate in the
z
z
-1
À4
À1
(
g ˆ 1.941; A ˆ 166 10 cm ), in both of which there is a cis
pH range 5 ±9. The practically constant EPR parameters
suggest the same environment around the metal ion (only the
protonation state of the phosphate is different) as was found
z
z
[27, 28]
arrangement of the ligand functions.
This strongly suggests
À
2‡
that [VOAB] in the VO -5MeOpicH system also features an
À
À
[28]
equatorial {Naryl, CO , CO , H O} donor set.
for the corresponding complexes with pic. Comparison of
2
2
2
Lactate forms the ternary complexes [VOAB] and
the pK(VOABHx) (x ˆ 2, 1) values for 5MeOpic (3.55 and
6.17) with those in the pic system (4.24 and 7.21) shows that
the stepwise deprotonation of the mono-coordinated phos-
phate takes place at lower pH values for 5MeOpic, indicating
somewhat stronger interaction between the metal ion and the
ligand.
À
[
4
VOABH ] , which are the major species in the pH range
À1
.5 ±7.5. The overall situation compares to that of the
corresponding pic and 6Mepic systems, for which deprotona-
tion of the alcoholic group of lactate was suggested for
À
[31]
[
VOABH ] . Since the EPR parameters do not change
À1
significantly on deprotonation, the alcoholic group is suppos-
edly in the axial position. The equatorial ligand set for both
The speciation results indicate that, after absorption and
transport into the blood, [VO(5MeOpic) (H O)] (2a) under-
2
2
4930
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4924 ±4935

Insulin-Mimetic Vanadium Complexes
4924±4935
goes considerable transformations when it meets the more
important low molecular mass bio-ligands of the blood serum.
This is illustrated by the data in Table 8, which provides the
distribution of VO2 among picolinate and the four serum
then added, which stains living cells only. At concentrations of
about 1 mm, the V complexes caused morphological changes
IV
up to total lysis, while at concentrations of 0.1 mm and below,
‡
V
the compounds were non-toxic. The V picolinates were non-
toxic even at 1 mm concentrations.
Table 8. Species distribution of insulin-mimicking VO2‡ picolinato complexes;
The insulin-mimetic tests on 2a ± c, K[VO (pic) ] and
2 2
calculated for c(V) ˆ 10 mM, phosphate 1.10 mm, citrate 99.0 mm, lactate
K[VO (dipic)] with respect to uptake and degradation of
2
1
.51 mm, and oxalate 9.20 mm; at pH 7.4.
glucose were performed by the MTTreduction essay. Soluble,
yellow MTT is reduced in the last part of the mitochondrial
electron transfer chain in living cells in to insoluble fomazan
blue. Cells were incubated in an insulin-free medium for 24 h,
followed by 3 h incubation with solutions of the vanadium
compound. The amount of farmazan blue formed, determined
photometrically, is proportional to the amount of glucose
taken into and degraded within the cells. The results are
represented in Figure 6. All of the compounds are effective at
subtoxic concentrations; 1a, K[VO (pic) ] and K[VO (dipic)]
VO2 in
‡
% VO bound
2‡
[VO(5MeOpic)2] [VO(6Mepic)2] [VO(pic)2]
intact [VO(picolinate)
ternary complexes with
citrate
2
]
0
0
0
27
14
30
0
42
0
phosphate
binary complexes with
citrate
lactate
50
9
61
9
50
8
2
2
2
are comparable to insulin itself.
ligands. The modelling calculations were performed at
The insulin-mimetic activity of ligand 1a and complex 2a
was also studied with respect to the in vitro inhibition of the
release of free fatty acids (FFA) from isolated rat adipocytes
treated with epinephrine (adrenalin). Epinephrine activates
lipases; the action of lipases in turn is counteracted by insulin.
The results, presented in Figure 7, are compared with those of
2‡
c(VO ) ˆ 10 mm (this is a concentration level at which 2a
still shows a marked insulin-mimetic effect; see below), while
the genuine serum concentrations were used for the bio-
ligands. These findings indicate that both picolinato ligands
are partly or completely displaced from the coordination
2‡
‡
sphere of VO , mostly so by citrate, which binds 70 ±90% of
the total VO2 in binary or ternary complexes. Phosphate and
lactate play a secondary role, and oxalate cannot compete.
Preliminary calculations of the species distribution including
VOSO ([VO(OH)(H O) ] ), VS, which is commonly used as
4
2
4
‡
[34]
a positive control and standard in these tests, and [VO(-
H O)(pic) ]. The inhibition was measured by the IC values,
2
2
50
that is 50% inhibition concentration of a compound for the
FFA release from adipocytes. Figure 7 shows the effect of 1a,
2
‡
the transport protein transferrin (Tf), using the VO -Tf
[
32]
binding constants logK ˆ 13.2 Æ 0.8 and logK ˆ 12.2 Æ 0.8
2a, VS and [VO(pic) ]. Since 2a exhibited a dose-dependent
1
2
2
(
where K was derived from the linear correlation between
inhibitory effect as is the case with VS and [VO(pic) ], while
1
2
the first Tf binding constant and the first hydroxide binding
the ligand 1a alone showed no such effect, the apparent IC50
constant of the metal ion,[33] and K is based on the fact that
values (Table 9) were rated by comparison with VS (IC ˆ
2
50
the second Tf binding constant is usually smaller than the first
one by one order of magnitude) show that about 65% of the
vanadyl forms a binary complex with Tf, while the remaining
approximately 35% are distributed between ternary vanadyl-
1.0 mm). On the basis of IC values, complex 2a has a
5
0
substantially higher insulin-mimetic activity than VS. The
difference in activities of 2a and [VO(pic) ] was within the
2
experimental error range, and also compares to that of the
[11]
5
MeOpic-citrate and -phosphate species.
iodo derivative [VO(5Ipic) (H O)]. The ligand 1a itself is
2 2
inactive.
Insulin-mimetic and cytotoxici-
ty tests: The biological tests for
the determination of the insu-
lin-mimetic abilities and the
short term cell toxicity of 2a ±
d were carried out on mice
fibroblasts, transformed so as
to exhibit a metabolism compa-
rable to that of adipocytes.[
K[VO (pic) ] and K[VO (dip-
4]
2
2
2
ic)], the insulinmimetic behav-
iour of which has been estab-
lished earlier (see Introduc-
tion), were included in these
investigations.
For cytotoxicity tests, the fi-
broblasts were incubated with
solutions of the vanadium com-
plexes at different concentra-
tions for 3 h. Trypan blue was
Figure 6. Glucose uptake stimulated by vanadium-picolinates. The absorbance is a measure of the efficacy of
the compounds. The diagram also contains the data for a control group (neither insulin nor vanadium
compound added), and for a group where insulin was employed instead of vanadium.
Chem. Eur. J. 2003, 9, 4924 ±4935
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4931

D. Rehder et al.
FULL PAPER
compound 2a, which is clearly
more effective than 2b and 2c
in triggering intracellular glu-
cose degradation (Figure 6), is
also the one which is most
effectively absorbed. The ab-
sorption is also better by a
factor of two than for the
anionic picolinato complexes
and vanadate.
Conclusion
Figure 7. Inhibitory effect of vanadium complexes on fatty acid release from rat adipocytes treated with
À5
epinephrine (10 m; 3 h). Data are expressed as the mean ÆSDs for three experiments. The blank corresponds to
Two of the major functions of
insulin are the activation of
glucose uptake by glucose-me-
tabolising cells from the blood
adipocytes without any treatment, the control to adipocytes pre-incubated with saline for 30 min. * and **:
significance at p < 0.05 and 0.01, respectively.
Table 9. IC50 values
serum, and the inhibition of lipolysis. Lacking insulin supply
diabetes 1) or insulin resistance (diabetes 2) leads to intol-
(
Compound
IC50 (mm)
SD
erable blood glucose levels and the formation of ketonic
bodies which cause a variety of symptoms such as a
malfunction of the peripheral blood circulation. Here we
introduce a novel family of picolinatovanadium compounds
for potential oral treatment of both types of diabetes mellitus.
These compounds contain 5-carboalkoxypicolinates, which
allow for a fine-tuning of the stability of the complexes and
their lipophilicity/hydrophilicity. The complexes retain their
identity under the acidic conditions encountered in the
stomach. Under pH conditions pertinent to the blood (7.35),
mainly vanadyl hydroxides form. However, in the presence of
the biogenic blood constituents citrate and phosphate (and, to
a certain extent, also oxalate and lactate), stable ternary
complexes form that still contain the specifically designed
picolinate derivative and should be considered the active
forms of the vanadium complexes in insulin mimesis. In
VOSO
4
1.00
none
0.56
0.48
0.16
1
2
a
a
0.11
0.08
[
2
VO(pic) ]
Uptake of vanadium by fibroblasts: To assess the biological
efficacy of a compound, its uptake by cells is an important
feature. We have incubated SVT fibroblasts with 1m and 0.1m
solutions of the compounds employed in the insulin-mimetic
glucose uptake tests (2a ± c, K[VO (pic) ], K[VO (dipic)],
2
2
2
including vanadate which is usually considered to penetrate
cell walls easily through phosphate channels) and determined
the cell vanadium contents by AAS after removal of the
supernatant solutions and thorough washing of the cells. The
results, collated in Table 10, have also been related to the cell
mass as quantified by protein, determined by the BCA
method (see Experimental Section). The data show that 3 ±
particular, the complex of the methyl ester, [VO(H O)(5-
2
nÀ
MeOpic) ] (2a)–or essentially [VO(5MeOpic)citrate] un-
2
15% of the vanadium in the original solution gets in the cells.
der physiological conditions–effectively mimics insulin in
vitro with respect to glucose uptake by mice fibroblasts and
inhibition of lipolysis (release of free fatty acids) by rat
adipocytes. 2a is more effectively taken up by fibroblasts than
the more lipophilic ethyl and isopropyl ester compounds (2b
and 2c), and the more hydrophilic anionic picolinato com-
plexes. 2a is thus expected to exhibit in vivo normoglycemic
activity.
The concentrations, in the micromolar range, are, however,
still on a level relevant for physiological action. Interestingly,
Table 10. Vanadium uptake by SVT mice fibroblasts.[a]
ng Vmg protein[b]
À1
Compound
mg Vanadium per L
concentration, mM)
(
2
2
2
2
2
2
a, 1 mm
757 (14.8)
73 (1.4)
135 (2.6)
21 (0.4)
362 (7.1)
37 (0.7)
431 (8.5)
42 (0.8)
373 (7.3)
68 (1.3)
373 (7.3)
68(1.3)
1.67
0.12
0.28
0.03
0.73
0.06
0.79
0.07
0.76
0.14
0.76
0.14
a, 0.1 mm
b, 0.1 mm
b, 0.1 mm
c, 0.1 mm
c, 0.1 mm
Experimental Section
Instrumentation: Infrared spectra were recorded as KBr pellets on a Perkin
Elmer 1720 FTIR spectrometer and NMR spectra were recorded on a
Varian Gemini 200 BB spectrometer at room temperature, relative to TMS
K[VO
K[VO
K[VO
K[VO
K[VO
K[VO
2
2
2
2
3
3
(pic)
(pic)
(dipic)], 1 mm
2
2
], 1 mm
], 0.1 mm
1
13
51
for H (200 MHz) and C (50 MHz), and relative to external VOCl
3
for V
(
53 MHz). EPR spectra were run in the X-band mode (9.75 MHz) on a
(dipic)], 0.1 mm
[
c]
BrukerESP 300E spectrometer. Simulations were carried out with the
program system SimFonia. Mass spectrometry was carried out with the
usual spectrometer settings (70 eV, EI/FAB) on a Varian MAT 311A
(70 eV, EI) or a VG Analytical 70-250S (FAB) instrument. The thermog-
ravimetric analysis of 2a was carried out with an STA 409 instrument
(Netzsch), coupled to a quadrupole mass spectrometer QMG 421 (Balzers)
] , 1 mm
[
c]
] , 0.1 mm
À1
[
a] Control <10 mgVL . [b] The average value of cell protein is 554 mgml.
À
[
c] Present in solution as an equilibrium mixture mainly of H
2
VO
4
,
2
À
4À
H
2
V
2
O
7
and V
4
O
10
.
4932
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4924 ±4935

Insulin-Mimetic Vanadium Complexes
4924±4935
À1
p
q
r
s
with
a
heating program of 5 Kmin between 30 and 8508C. pH-
[M
PSEQUAD computer program.
stability constants are given in parentheses in the tables. During the
p q r s
H A B ]/[M] [H] [A] [B] for ternary species were calculated with the
[
37]
Potentiometric measurements were performed with
pHM 84 equipped with Metrohm combined electrode (type
.0234.110), calibrated for hydrogen ion concentration according to
a
radiometer
The uncertainties (3s values) of the
a
2
‡
6
calculations the following VO -hydroxo complexes were assumed:
[
36]
‡
2‡
Irving.
[VO(OH)] (log1,À1,0 ˆ À5.94), [{VO(OH)}
2
]
(log2,À2,0 ˆ À6.95), calculat-
[38]
ed from the data published by Henry et al., the Davie×s equation being
X-ray structure analyses were carried out at 153(2) K using MoKa radiation
À
used to take into consideration the different ionic strengths, [VO(OH)
3
]
(
l ˆ 0.71073 ä, graphite monochromator) on a Smart Apex CCD diffrac-
À
(
log1,À3,0 ˆ À18.0) and [(VO)
2
(OH)
5
]
(log2,À5,0 ˆ 22.5) taken from ref. [39].
tometer at 153(2) K. Hydrogen atoms were usually found and refined
anisotropically, and otherwise calculated into idealised positions and
included in the last cycles of refinement. Absorption corrections were
The ionic strength was adjusted to 0.20m KCl in each solution studied. In all
cases, the temperature was maintained at 25.0 Æ 0.18C.
carried out by SADABS. Disorder information: 2a ¥ 3.5H
2
O: In one of the
Toxicity and insulin-mimetic tests: Cytotoxicity and glucose uptake tests
were performed on Simian virus modified Swiss 3T3 mice fibroblasts (cell
line SV 3T3), maintained as described elsewhere.[ SV modified fibroblasts
attain metabolic properties closely related to adipocytes and are thus
particularly suited in studies of glucose uptake and metabolism. For the
tests, cells were grown to sub-confluency. For toxicity tests, the cells, kept in
Dulbecco×s modification of Eagle×s medium (DMEM, Sigma), were
incubated with solutions of the vanadium compounds for 3 h. The
supernatant medium was removed. Trypan blue (0.2% w/v) in phos-
phate-buffered saline solution was added, and the ratio of stained and non-
stained cells determined after 5 min. Tests for the glucose uptake activity
were based on the MTTreduction essay. We have shown previously that the
results obtained by the MTT test are compatible with direct measurements
two independent molecules, the carbomethoxy group (C32O19) disordered
over two positions was treated with a 0.7:0.3 model. Four of the seven water
molecules of crystallisation showed partial occupancy/were disordered and
were dealt with by employing the following REM: O24 0.5, O25/O26 0.75/
.25, O27/O28 0.65/0.35, O29/O30 0.3/0.2, O31/O32/O33/O35 0.35/0.30/
.20/0.15. 3: Disorder of 1H
O per 0.5 V; REM ˆ O11 0.18, O12 0.12, O13
.06, O14 0.15, O15 0.05, O16 0.10, O17 0.15, O18 0.07, O19 0.05, O20 0.07.
For crystal data and structure refinement see Table 11.
CCDC-207528(1b), CCDC-207529 (1a ¥ H O), CCDC-207530 (1c),
CCDC-207531 ([NH ]3 ¥ 4H O), CCDC-207532 (2a ¥ 3.5H O) contain the
supplementary crystallographic data for this paper. These data can be
obtained free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or
from the Cambridge Crystallographic Data Centre, 12, Union Road,
Cambridge CB21EZ, UK; fax: (‡44)1223-336-033; or deposit@ccdc.cam.
ac.uk).
4]
0
0
0
2
2
4
2
2
[
4]
of the uptake of tritium labelled glucose. The cells were incubated in
insulin-free, supplemented[ DMEM medium for 24 h. DMEM (without
phenol red) supplemented with MTT (Sigma; 0.5 gL ) and the solution of
4]
À1
the vanadium complex were added to the cells and incubated for 3 h at
Potentiometric measurements: The stability constants of the proton and
VO2‡ complexes of the ligands were determined by pH-potentiometric
378C in a 5% CO atmosphere. The reaction was then stopped by addition
2
of 0.05m HCl in 2-propanol. The dye was extracted with HCl/2-propanol,
and the amount of formazan blue formed by reduction of MTT was
measured at 570 nm in a multi-well reader (SLT 340 ATC).
titration of 20.0 mL samples. The picolinic acid (5MeOpicH, A) concen-
2‡
tration was in the range 2 ±6 m m, and the molar ratios in the VO -A
system were 0:1, 1:1, 1:2, 1:4. For the ternary systems containing lactate,
2
‡
oxalate, phosphate or citrate, B, the VO :A:B ratios were 1:1:1, 1:1:2,
:2:1, or 1:2:2. All of the titrations were performed over the pH range 2 ±11,
For determination of the vanadium uptake by the fibroblasts, the cells were
pre-treated accordingly (3 h of incubation with the solutions of the
vanadium complexes) on micro-well plates. The supernatant was discarded,
the cells washed twice with buffer, treated with 1 mL of 0.5m NaOH per
well and stirred over night to ascertain complete digestion. Half of each
sample was then subjected to AAS (graphite tube) on a PE 4100ZL
instrument. For the analyses, 20 mL sample volumes were employed,
1
or until precipitation occurred, with a carbonate-free KOH solution of
known concentration (ꢀ0.2m), under a purified argon atmosphere. The
reproducibility of the titration points included in the evaluation was within
0
.005 pH units over the whole pH range. A pK
w
value of 13.76 Æ 0.01 was
determined and used for the calculations. The concentration stability
constants bpqr ˆ [M
p
q
r
p
H
q
A
r
]/[M] [H] [A] for binary complexes and bpqrs
ˆ
acidified with HNO
3
and 5 mg Pd ‡ 3 mg MgNO
3
added as modifier.
Table 11. Crystal data and structure refinement for 1a ¥ H
a ¥ H
NO
2
O, 1b, 1c, 2a ¥ 3.5H
2
O and [NH
4
2
]-3 ¥ 4H O.
1
2
O
1b
1c
2a ¥ 3.5H
2
O
4 2
[NH ]-3 ¥ 4H O
empirical formula
C
8
H
9
5
C
9
H
9
NO
4
C
10
H
11NO
4
16 24 3 14
C H N O V
533.32
À1
M [gmol
]
199.16
195.17
209.20
crystal system
space group
cell dimensions
a [ä]
b [ä]
c [ä]
triclinic
P1
triclinic
P1
monoclinic
P2(1)/c
triclinic
P1
monoclinic
C2/c
≈
≈
≈
7.1533(12)
7.2375(12)
9.5728(16)
81.499(3)
76.958(2)
65.886(2)
439.82(13)/2
1.504
9.4412(3)
10.8428(4)
13.4610(5)
102.5600(10)
97.8760(10)
97.5560(10)
1313.65(8)/6
1.480
12.3015(4)
9.2940(3)
8.9091(3)
90
95.4080(10)
90
1014.04(6)/4
1.370
0.107
13.0348(8)
13.3654(8)
15.1968(10)
104.607(1)
100.850(1)
116.090(1)
2156.4(2)/4
1.568
18.121(3)
15.169(2)
10.0449(15)
a [8]
b [8]
119.233(2)
g [8]
V [[ä] /Z]
3
2409.5(6)/4
1.470
0.484
À3
1
calcd [gcm
]
À1
m [mm
F(000)
]
0.127
208
0.118
612
0.534
1043
440
1104
crystal size [mm]
q range [8]
index ranges
0.75 Â 0.20 Â 0.05
0.80 Â 0.70 Â 0.30
2.19 ±32.00
À 13 < h < 18
À 16 < k < 16
À 13 < l < 13
34668
8881 (0.0550)
0/388
1.020
0.65 Â 0.53 Â 0.34
0.05 Â 0.05 Â 0.15
1.48 ±25.02
À 23 < h < 23
À 15 < k < 15
À 18 < l < 18
21985
7577 (0.0881)
1/599
0.840
0.43 Â 0.31 Â 0.17
2.19 ±27.53
2.75 ±31.99
2.44 ±27.49
À 9 < h < 9 À14 < h < 14
À 9 < k < 9
À 20 < l < 20
5058
1973 (0.0268)
0/140
À 15 < h < 15
À 11 < k < 13
À 19 < k < 19
À 13 < l < 13
14245
2750 (0.0302)
4/190
À 12 < l < 12
reflections collected
independent reflections (Rint
restraints/parameters
GooF on F 2
14255
3457 (0.0274)
0/140
)
1.019
1.089
1.038
final R (I > 2sI
R indices (all data), R1/wR2
largest diff. peak and hole [eä
0
), R1/wR2
0.0417/0.1116
0.0488/0.1154
0.236/ À 0.0229
0.0600/0.1524
0.0770/0.1619
0.680/ À 0.429
0.0491/0.1445
0.0555/0.1481
0.490/ À 0.202
0.0561/0.1325
0.1348/0.1814
0.624/ À 0.482
0.0445/0.1275
0.0533/0.1317
0.516/ À 0.264
À3
]
Chem. Eur. J. 2003, 9, 4924 ±4935
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4933

D. Rehder et al.
FULL PAPER
Calibration was carried out against a 0.5m NaOH solution containing
([D
6
]DMSO): d ˆ 165.48 (C-2'), 164.07 (C-5'), 151.66 (C-2), 149.75 (C-6),
1
00 mg V and 500 mg BSA (bovine serum albumin) per L. The second half
2 3
138.24 (C-4), 128.10 (C-5), 124.62 (C-3), 61.65 (CH CH ), 14.16 ppm
of each sample of the digested cell material was subjected to protein
determination by the BCA (bicinechoninic acid) method on 10 mL samples.
Calibration was carried out against 10 mL samples of a BSA dilution
sequence.
(CH
2 3 9 9 4
CH ); elemental analysis calcd (%) for C H NO (195.17): C 55.39, H
7.18, N 4.65; found: C 55.27, H 7.16, N 4.55; MS (70 eV, EI): m/z (%): 195 (2)
‡
.
[M ], 151 (100) [M À CO
2
], 150 (79) [M À C
N].
2
H
5
O], 123 (37) [M À C
3 4 2
H O ],
106 (12) [M À C ], 78 (14) [C
3
H
5
O
3
5 4
H
Tests for the inhibition of free fatty acid (FFA) release were performed on
rat adipocytes. Male Wistar rats were killed by bleeding under anaesthesia
with ether (animal experiments were approved by the Experimental
Animal Research Committee of Kyoto Pharmaceutical University and
performed according to the guidelines), and adipose tissues were removed,
chopped, and digested with collagenase for 1 h at 378C in Krebs-Ringer
bicarbonate buffer, pH 7.4, containing 2% bovine serum albumin. The
adipocytes thus obtained were separated from undigested tissue by
filtration through nylon mesh, and washed three times with collagenase
free buffer. 15 mL of saline glucose solution and 30 mL of the vanadium
5-Carboisopropoxypyridine-2-carboxylic acid, 5iPrOpicH (1c): IR (KBr):
nÄ ˆ 3053 (ar. C À H); 2980, 2875, 2767 (CH); 1718 (CˆO); 1600, 1583, 1498
(CˆC), (CˆN); 1459, 1471, 1377 (CH), (OH); 1301, 1269 (CˆC), (CˆN);
À1
1
1108, 1036 (CÀO); 750 cm (ar. CÀH); H NMR ([D
]DMSO): d ˆ 9.12
6
4
5
4
3
(dd, J6,4 ˆ 2.17 Hz, J6,3 ˆ 0.82 Hz, 1H; H-6), 8.40 (dd, J4,6 ˆ 2.17 Hz, J4,3
ˆ
5
3
8.10 Hz, 1H; H-4), 8.13 (dd, J3,6 ˆ 0.82 Hz, J3,4 ˆ 8.10 Hz, 1H; H-3), 5.16
3
3
(e, J ˆ 6.23 Hz, 1H; CH(CH
3
)
2
), 1.32 ppm (d, J ˆ 6.23, 6H; CH(CH
3 2
) );
1
3
C NMR ([D
6
]DMSO): d ˆ 166.17 (C-2'), 164.21 (C-5'), 152.32 (C-2),
150.39 (C-6), 138.82 (C-4), 128.98 (C-5), 125.24 (C-3), 69.95 (CH(CH
22.21 ppm (CH(CH ); elemental analysis calcd (%) for C10
(209.20): C 57.41, H 5.30, N 6.70; found: C 57.34, H 5.31, N 6.67; MS
(70 eV, EI): m/z (%): 209 (5) [M ], 168 (42) [M À C
CO ], 150 (100) [M À C O], 123 (41) [M À C ], 43 (21).
-Carbo(R,S)-secbutylmethoxypyridine-2-carboxylic acid, 5sBuOpicH
3 2
) ),
3
)
2
H11NO
4
6
À1
complexes dissolved in saline/DMSO were added to 240 mL (10 cellsmL
)
‡
.
of isolated adipocytes, and the resulting suspensions pre-incubated at 378C
for 30 min. After addition of 15 mL of 0.2 mm epinephrine dissolved in
saline, the resulting mixture was incubated (378C, 3 h). The reaction was
stopped by cooling with ice-water. The FFA concentration in the outer
solution of the cells was determined by an NEFA kit (Wako Pure
Chemicals). Final concentrations were as follows: glucose 5 mm; vanadium
compounds 0.1, 0.5 and 1 mm; ligand 1a 0.2, 1 and 2 mm; epinephrine
3
H
5
] , 165 (34) [M À
2
3
H
7
4 6 2
H O
5
(
(
(
(
1d): IR (KBr): nÄ ˆ 3047 (ar. C H); 2976, 2935, 2878, 2768 (CH); 1715
À
C O); 1599, 1581 (C C), (C N); 1457, 1419, 1372 (CH), (OH); 1297, 1267
ˆ
ˆ
ˆ
À1
1
C C), (C N); 1119, 1109, 1035 (C O); 749 cm (ar. C H); H NMR
ˆ
ˆ
À
À
3
[D
6
]DMSO, TMS): d ˆ 9.17 (br, 1H; H-6), 8.46 (d,
J
4,3 ˆ 8.06 Hz, 1H;
)CH CH ),
), 1.33 (d, J ˆ 6.23 Hz, 3H;
)CH CH );
0
.01 mm; DMSO 2% v/v.
Reagents and general: Vanadylsulfate VOSO
pyridin-2,5-dicarboxylic acid were obtained from Merck and Fluka and
used without further purification. K[VO (dipic)] was prepared as descri-
] was obtained in analogy to the respective
H-4), 8.19 (br, 1H; H-3), 5.06 (ps, J ˆ 6.23 Hz, 1H; CH(CH
3
2
3
¥ 5H
2
O, [NH
4
][VO
3
] and
1.78 ±1.64 (m, 2H; CH(CH
CH(CH )CH CH
C NMR ([D
148.88 (C-6), 138.26 (C-4), 128.26 (C-5), 124.75 (C-3), 73.63
(CH(CH )CH CH ), 28.26 (CH(CH )CH CH ), 19.21 (CH(CH )CH CH ),
9.53 ppm (CH(CH )CH CH ); elemental analysis calcd (%) for C11
3
)CH
2
CH
3
4
3
3
2
3
), 0.94 ppm (t, J ˆ 7.42 Hz, 3H; CH(CH
3
2
3
1
3
6
]DMSO/TMS): d ˆ 165.87 (C-2'), 163.84 (C-5'), 151.9 (C-2),
2
[
40]
2 2
bed, , and K[VO (pic)
[
18]
ammonium salt. Oxalate, lactate, citrate and phosphate were Aldrich
products of puriss. quality (ꢁ98.5%) and used as received. Their purities
were checked and the exact concentrations of the prepared stock solutions
were determined by the Gran method. VO stock solution was prepared
according to Nagyp a¬ l and F a¬ bi a¬ n[ and standardised for metal ion
concentration by permanganate titration and for hydrogen ion concen-
tration by pH-potentiometry, using the appropriate Gran function.
3
2
3
3
2
3
3
2
3
3
2
3
H
13NO
4
(223.23): C 59.19, H 5.87, N 6.27; found: C 58.88, H 5.91, N 6.31; MS (70 eV,
[41]
2‡
‡.
EI): m/z (%): 223 (1) [M ], 168 (60) [M À C
4
H
7
], 150 (100) [M À C
4 9
H O],
42]
122 (25) [M À C
General procedure for the preparation of the vanadium compounds
VO(H O)(5ROpic) ] (2a ±d): A solution of VOSO ¥ 5H O (127 mg,
0.5 mmol) in H O (2 mL) was added to a solution of 5ROpicH (1.0 mmol)
and NaO CCH ¥ 3H O (136 mg, 1.0 mmol) in water (10 mL) plus THF
5 9 2
H O ].
[
2
2
4
2
All syntheses with vanadium in oxidation state ‡iv were carried out under
2
an atmosphere of nitrogen using Schlenk techniques. Tetrahydrofuran
2
3
2
(
THF) was dried and deoxygenated by refluxing for 3 h over LiALH
distilled in an N stream. Deionised water was degassed before use.
Syntheses of compounds
4
, and
(1 mL). After 2 h of stirring at 508C, a green precipitate was formed which
was isolated by filtration and dried in vacuo. Yields ranged from 60 to 85%.
2
2 2
[VO(H O)(5MeOpic)
] (2a): IR (KBr): nÄ ˆ 3115, 3070 (ar. C
ÀH); 2958
À
(
(
(
CH
CO
3
); 1733 (CˆO); 1654 (CO
2
); 1578 (CˆC), (CˆN); 1439, 1397 (CH
3
),
General procedure for the 5-carboalkoxypicolinates 1a ± d (see also
Scheme 1): The syntheses of the dialkylesters from dipicolinic acid via
À
); 1303, 1286 (CˆC), (CˆN); 1124, 1045 (OÀCH
); 971 (VˆO); 833
2
3
À1
ar. CÀH), (CÀCO
); 750 cm (ar. CÀH); elemental analysis calcd (%) for
[
13]
2
the acid dichloride followed literature procedures.
Synthesis of the
C
16
H
14
N
2
O
10V (445.23): C 43.16, H 3.17, N 6.29 V 11.68; found: C 42.50, H
monoesters was carried out according to Ooi and Magee,[ with the
modification that a molar ratio diester to copper(ii)nitrate of 1:1 was chosen
and the reaction mixture kept at reflux for 3 h to yield the blue copper
14]
3
.30, N 6.27, V 11.44; MS (FAB): m/z (%): 428.
[
2
1
VO(H
2
O)(5EtOpic)
2
] (2b): IR (KBr): nÄ ˆ 3120, 3059 (ar. CÀH); 2980,
À
936, 2919 (CÀH); 1737 (CˆO); 1651 (CO
); 1610, 1574 (CˆC), (CˆN);
complexes [Cu(5ROpic)
The ligands were isolated as white or slightly yellow solids in yields of 44 to
7
2 2
], from which the ligands were liberated with H S.
2
À
489, 1402, 1364 (CH), (CO
2
); 1294 (CˆC), (CˆN); 1121, 1045 (CÀO); 978
À1
(
VˆO); 849 (ar. CÀH), (CÀCO
calcd (%) for C18 10V¥ 0.5H
found: C 45.03, H 3.98, N 5.53; MS (FAB): m/z (%): 456.
2
); 753 cm (ar. CÀH); elemental analysis
2%.
H
18
N
2
O
2
O (482.29): C 44.83, H 3.97, N 5.81;
5
-Carbomethoxypyridine-2-carboxylic acid, 5MeOpicH (1a): IR (KBr):
nÄ ˆ 3089, 3011 (ar. CÀH); 2967 (CH
); 1719 (CˆO); 1698 (CˆO); 1594,
3
570, 1484 (CˆC), (CˆN); 1421, 1383 (CH ), (OH); 1309, 1289, 1255 (CˆC),
[
VO(H
2
O)(5iPrOpic)
2
] (2c): IR (KBr): nÄ ˆ 3114, 3058 (ar. CÀH); 2983,
1
3
À
À1
1
2873 (CÀH); 1730 (CˆO); 1687, 1652 (CO ); 1609 (CˆC), (CˆN); 1484,
(
CˆN); 1115, 1020 (OÀCH
3
); 746 cm (ar. CÀH); H NMR ([D
6
]DMSO,
2
À
4
6,4 ˆ 1.89 Hz, 5J6,3 ˆ 0.85 Hz, 1H; H-6), 8.56 (dd,
1397, 1353 (CH), (CO ); 1289 (CˆC), (CˆN); 1108, 1051 (CÀO); 967
TMS): d ˆ 9.22 (dd,
J
2
À1
4
3
5
3
(VˆO); 848 (ar. CÀH), (CÀCO ); 750 cm (ar. CÀH); elemental analysis
J
4,6 ˆ 1.89 Hz,
J
4,3 ˆ 8.13 Hz, 1H; H-4), 8.32 (dd,
J
3,6 ˆ 0.85 Hz,
J
3,4
ˆ
2
1
3
20 22 2 10
calcd (%) for C H N O V (501.34): C 47.91, H 4.42, N 5.59; found: C
8
.13 Hz, 1H; H-3), 4.02 ppm (s, 3H; CH
d ˆ 165.40 (C-2'), 164.52 (C-5'), 151.62 (C-2), 149.70 (C-6), 138.24 (C-4),
27.82 (C-5), 124.54 (C-3), 52.68 ppm (OCH ); elemental analysis calcd
%) for C NO ¥ H O (199.16): C 48.25, H 4.55, N 7.03; found: C 48.21, H
.35, N 7.02; MS (70 eV, EI): m/z (%): 181 (2) [M ], 150 (49) [M À CH
37 (100) [M À CO ], 122 (30) [M À C ], 106 (16) [M À C ], 78
N].
3
); C NMR ([D
6
]DMSO, TMS):
4
7.56, H 4.45, N 5.08; MS (FAB): m/z (%): 484.
1
(
4
1
(
3
[VO(H O)(5sBuOpic) ] (2d): IR (KBr): nÄ ˆ 3118, 3058 (ar. CÀH); 2975,
2
2
À
H
7
4
2
2927 (CÀH); 1729 (CˆO); 1650 (CO ); 1576 (CˆC), (CˆN); 1485, 1458,
8
2
‡
.
À
O],
1403, 1358 (CH), (CO ); 1293 (CˆC), (CˆN); 1125, 1049 (CÀO); 967
3
2
À1
2
H
3
O
2
2
H
3
O
3
(VˆO); 851 (ar. CÀH), (CÀCO ); 750 cm (ar. CÀH); elemental analysis
2
2
17) [C
5
H
4
calcd (%) for C H N O V (529.40): C 49.91, H 4.95, N 5.29; found: C
2
2
26
2
10
4
9.64, H 5.10, N 5.29; MS (FAB): m/z (%): 512.
[NH ][VO (5MeOpic) ] (3): A solution of 5MeOpicH (182 mg, 1.0 mmol)
in acetone (1 mL) was added to a slurry of [NH ][VO ] (59 mg, 0.5 mmol)
in water (0.5 mL). The resulting yellow solution was stirred overnight and
kept at 48C for two days. Yellow single crystals of [NH ]-3 ¥ 4H O were
5
-Carboethoxypyridine-2-carboxylic acid, 5EtOpicH (1b): IR (KBr): nÄ ˆ
3
131 (ar. CÀH); 2984, 2943, 2821 (CH); 1718 (CˆO); 1601, 1577 (CˆC),
4
2
2
(
(
CˆN); 1476, 1440, 1383, 1371 (CH), (OH); 1302, 1279, 1247 (CˆC),
4
3
À1
1
CˆN); 1118, 1028 (CÀO); 748 cm (ar. CÀH); H NMR ([D
]DMSO):
6
5
4
3
d ˆ 9.15 (d,
J
6,3 ˆ 1.28 Hz, 1H; H-6), 8.44 (dd,
J
4,6 ˆ 2.01 Hz,
J
4,3
ˆ
4
2
3
3
8
2
.06 Hz, 1H; H-4), 8.15 (d, J3,4 ˆ 8.06 Hz, 1H; H-3), 4.37 (q, J ˆ 7.11 Hz,
obtained from the reaction solution by further cooling. Elemental analyses
were not carried out due to very low yields of pure, crystalline material. IR
3
13
H; CH
2
CH
3
), 1.34 ppm (t, J ˆ 7.11 Hz, 3H; CH
2
CH
3
);
C NMR
4934
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4924 ±4935

Insulin-Mimetic Vanadium Complexes
4924±4935
À
(
1
KBr): nÄ ˆ 3114, 3081 (ar. CÀH); 2963 (CH
3
); 1732 (CˆO); 1658 (CO
2
);
[18] M. Melchior, K. H. Thompson, J. M. Jong, S. J. Rettig, E. Shuter, V. G.
À
580 (CˆC), (CˆN); 1439, 1396 (CH
), (CO
); 1348, 1304, 1283 (CˆC),
Yuen, Y. Zhou, J. H. McNeill, C. Orvig, Inorg. Chem. 1999, 38, 2288 ±
2293.
[19] D. C. Crans, L. Yang, T. Jakusch, T. Kiss, Inorg. Chem. 2000, 39, 4109 ±
3
2
(
7
CˆN); 1125, 1037 (OÀCH
3
); 903 (VˆO); 836 (ar. CÀH), (CÀCO
O): d ˆ 9.05 ±7.95 (m, 3H; H-6, H-4,
); V NMR (D
O): d ˆ À575 ppm.
2
);
À1
1
51 cm (ar. CÀH); H NMR (D
2
5
1
H-3), 3.95 ppm (s, 3H; CH
3
2
4416.
[
20] D. S. Greco, Diabetes 1997, 46, 1289.
[
21] A. Shaver, J. B. Ng, D. A. Hall, B. S. Lum, B. I. Posner, Inorg. Chem.
1
993, 32, 3109 ±3113.
Acknowledgement
[
22] L. Yang, D. C. Crans, S. M. Miller, A. La Cour, O. P. Anderson, P. M.
Kaszynski, L. D. Austin, G. R. Willsky, J. Am. Chem. Soc. 2002, 41,
This work was supported by the Deutsche Forschungsgemeinschaft, the
Fonds der Chemischen Industrie, the European COST Programme D21-
4
859 ±4871.
[
[
[
23] T. H. Fife, T. J. Przystas, J. Am. Chem. Soc. 1982, 104, 2225 ±2257.
24] G. K. S. Ooi, R. J. Magee, J. Inorg. Nucl. Chem. 1970, 32, 3315 ±3320.
25] a) M. R. Maurya, S. Khurana, Shailendra A. Azam, W. Zhang, D.
Rehder, Eur. J. Inorg. Chem. 2003, 1966 ±1973; b) M. R. Maurya, S.
Khurana, W. Zhang, D. Rehder, J. Chem. Soc. Dalton Trans. 2002,
0
009-01 and the Hungarian Scientific Fund (OTKA T31896, F32235). P.B.
gratefully acknowledgements the J a¬ nos B o¬ lyai research fellowship for
financial support. We also thank Marion Eternot (SOKRATES student
from the Universit e¬ e Pierre et Marie Curie, Paris) and A.-L¸der Gaulke,
University of Hamburg, for support with the syntheses.
3
015 ±3023.
[
[
[
26] N. D. Chasteen, Biological Magnetic Resonance, vol. 3 (Eds.: L. J.
Berliner, J. Reuben), Plenum, New York, 1981.
27] T. Kiss, E. Kiss, E. Garriba, H. Sakurai, J. Inorg. Biochem. 2000, 80,
[
[
1] B. Lyonnet, X. Mertz, E. Martin, Presse Med. 1899, 1, 191.
2] K. H. Thompson, J. H. McNeill, C. Orvig, Chem. Rev. 1999, 99, 2561 ±
6
5 ±73.
28] E. Kiss, E. Garribba, G. Micera, T. Kiss, H. Sakurai, J. Inorg. Biochem.
000, 78, 97 ±108.
2
571.
[
[
3] H. Sakurai, Y. Kojima, Y. Yoshikawa, K. Kawabe, H. Yasui, Coord.
Chem. Rev. 2002, 226, 187 ±198.
2
[
[
29] W. R. Harris, J. Clin. Chem. 1992, 38, 1809 ±1818.
30] P. Bugly o¬ , E. Kiss, I. F a¬ bi a¬ n, T. Kiss, D. Sanna, E. Garribba, G. Micera,
Inorg. Chim. Acta 2000, 306, 174 ±183.
4] D. Rehder, J. Costa Pessoa, C. F. G. C. Geraldes, M. M. C. A. Castro,
T. Kabanos, T. Kiss, B. Meier, G. Micera, L. Pettersson, M. Ranger, A.
Salifoglou, I. Turel, D. Wang, J. Biol. Inorg. Chem. 2002, 7, 384–396.
5] D. C. Crans, J. Inorg. Biochem. 2000, 80, 123 ±131 and references
therein.
[
[
[
[
31] T. Kiss, P. Bugly o¬ , D. Sanna, G. Micera, P. Decock, D. Dewaele, Inorg.
[
[
[
Chim. Acta 1995, 239, 145 ±153.
32] T. Kiss, Abstracts of the 6th European Conference of Bioinorganic
Chemistry, Lund and Copenhagen, 2002, p. 48.
33] H. Sun, M. C. Cox, H. Li, P. Sadler, Struct. Bonding (Berlin), 1997, 88,
6] M. Melchior, S. J. Rettig, B. D. Liboiron, K. H. Thompson, V. G. Yuen,
J. H. McNeill, C. Orvig, Inorg. Chem. 2001, 40, 4686 ±4690.
7] P. J. Stankiewicz, A. S. Tracey, D. C. Crans, Vanadium and Its Role in
Life (Eds.: H. Sigel, A. Sigel), Metal Ions in Biological Systems,
vol. 31, Marcel Dekker, New York 1995, ch. 9.
7
1.
34] M. Nakai, H. Watanabe, C. Fujiwara, H. Kakegawa, T. Satoh, T.
Takada, R. Matsushita, H. Sakurai, Biol. Pharm. Bull. 1995, 18, 719 ±
7
[
[
8] J. Selbin, Coord. Chem. Rev. 1966, 1, 293 ±314.
9] H. Sakurai, K. Fujii, H. Watanabe, H. Tamura, Biochem. Biophys. Res.
Commun. 1995, 214, 1095 ±1101.
25.
35] K. Elvingson, A. Gonz a¬ les-Bar o¬ , L. Pettersson, Inorg. Chem. 1996, 35,
388 ±3393.
36] H. M. Irving, M. G. Miles, L. D. Pettit, Anal. Chim. Acta, 1967, 38,
75 ±488.
[
[
[
3
[
[
[
[
[
[
[
[
10] S. Fujimoto, K. Fujii, H. Yasui, R. Matsushita, J. Takada, H. Sakurai, J.
Clin. Biochem. Nutr. 1997, 23, 113 ±129.
4
11] T. Takino, H. Yasui, A. Yoshitake, Y. Hamajima, R. Matsushita, J.
Takada, H. Sakurai, J. Biol. Inorg. Chem. 2001, 6, 133 ±142.
12] G. R. Willsky, A. B. Goldfine, P. J. Kostyniak, J. H. McNeill, L. Yaang,
H. R. Khan, D. C. Crans, J. Inorg. Biochem. 2001, 85, 33 ±42.
13] J. V. Fondacaro, D. S. Greco, D. C. Crans, Annu. Vet. Med. Forum
37] L. Z e¬ k a¬ ny, I. Nagyp a¬ l in Computational Methods for the Determi-
nation of Stability Constants (Ed.: D. Leggett), Plenum, New York,
1
985.
38] R. P. Henry, P. C. H. Mitchell, J. E. Prue, J. Chem. Soc. Dalton Trans.
973, 1156 ±1159.
39] A. Komura, M. Hayashi, H. Imanaga, Bull. Chem. Soc. Jpn. 1977,
927 ±2931.
[
[
1
1
999, 710.
14] T. Sasagawa, Y. Yoshikawa, K. Kawabe, H. Sakurai, Y. Kojima, J.
Inorg. Biochem. 2002, 88, 108 ±112.
15] H. Yasui, A. Tamura, T. Takino, H. Sakurai, J. Inorg. Biochem. 2002,
1, 327 ±338.
16] L. Yang, A. La Cour, O. P. Anderson, D. C. Crans, Inorg. Chem. 2002,
1, 6322 ±6331.
2
[
[
[
40] K. Wieghardt, Inorg. Chem. 1978, 17, 57 ±64.
41] G. Gran, Acta Chem. Scand. 1950, 4, 559 ±577.
42] I. Nagyp a¬ l, I. F a¬ bi a¬ n, Inorg. Chim. Acta 1982, 61, 109 ±113.
9
4
17] D. C. Crans, L. Yang, J. A. Alfano, L.-H. Chi, W. Jin, M. Mahroof-
Tahir, K. Robbins, M. M. Toloue, L. K. Chan, A. J. Plante, R. Z.
Grayson, G. R. Willsky, Coord. Chem. Rev. 2003, 237, 13 ±22.
Received: April 7, 2003 [F5019]
Chem. Eur. J. 2003, 9, 4924 ±4935
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4935