Crystal structures and physico-chemical properties of Zn(II) and Co(II) tetraaqua(3-nitro-4-hydroxybenzoato) complexes: Their anticonvulsant activities as well as related (5-nitrosalicylato)-metal complexes

Article abstract of DOI:10.1016/j.poly.2007.10.006

Purposes of these studies were to synthesize Zn(II) and Co(II) complexes of 3-nitro-4-hydroxybenzoic acid, determine their structures through X-ray crystallography, and obtain their anticonvulsant activities. Thermogravimetric, differential scanning calor

Full text of DOI:10.1016/j.poly.2007.10.006

Available online at www.sciencedirect.com
Polyhedron 27 (2008) 537–546
www.elsevier.com/locate/poly
Crystal structures and physico-chemical properties of Zn(II) and Co(II)
tetraaqua(3-nitro-4-hydroxybenzoato) complexes: Their anticonvulsant
activities as well as related (5-nitrosalicylato)–metal complexes
a,
a,
a
Jean d’Angelo *, Georges Morgant *, Nour Eddine Ghermani b,c, Didier Desmaele ,
¨
c
b
d
d
e
Bernard Fraisse , Franc¸ois Bonhomme , Emma Dichi , Mehrez Sghaier , Yanling Li ,
e
f,
*
Yves Journaux , John R.J. Sorenson
a
ˆ
Laboratoire BioCIS, UMR CNRS 8076, IFR 141, Faculte´ de Pharmacie, Universite´ Paris-Sud, 5 rue J.-B. Cle´ment, 92296 Chatenay-Malabry, France
Laboratoire PPB, UMR CNRS 8612, IFR 141, Faculte´ de Pharmacie, Universite´ Paris-Sud, 5 rue J.-B. Cle´ment, 92296 Chaˆtenay-Malabry, France
b
c
ˆ
Laboratoire SPMS, UMR CNRS 8580, Ecole Centrale Paris, Grande Voie des Vignes, 92295 Chatenay-Malabry, France
d
ˆ
Laboratoire CPMB, Faculte´ de Pharmacie, Universite´ Paris-Sud, 5 rue J.-B. Cle´ment, 92296 Chatenay-Malabry, France
Laboratoire CIMM, UMR CNRS 7071, Universite´ Pierre et Marie Curie Paris 6, 4 Place Jussieu, 75005 Paris, France
e
f
Division of Medicinal Chemistry, Department of Pharmaceutical Sciences, University of Arkansas, Medical Sciences Campus, Slot 522,
4301 West Markham Street, Little Rock, AR 72205, USA
Received 4 September 2007; accepted 4 October 2007
Available online 14 November 2007
Abstract
Purposes of these studies were to synthesize Zn(II) and Co(II) complexes of 3-nitro-4-hydroxybenzoic acid, determine their structures
through X-ray crystallography, and obtain their anticonvulsant activities. Thermogravimetric, differential scanning calorimetry, imped-
ance of aqueous solutions and magnetic properties analyses were also determined. Anticonvulsant and related activities of these com-
plexes as well as Zn(II), Co(II), Ni(II) and Mg(II) (5-nitrosalicylato) complexes were determined by the National Institutes of Health,
Antiepileptic Development Program. Results of these analyses are presented to document unique bonding features and physical prop-
erties of these compounds and their anticonvulsant activities. It is concluded that these compounds have chemical and physical properties
that can be used to account for their anticonvulsant activities.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Keywords: Crystal structures; 3-Nitro-4-hydroxybenzoic acid ligand; Zn(II) and Co(II) complexes; Anticonvulsant activity
1. Introduction
[4–10]. As part of a systematic program devoted to the
study of structure–activity relationships of (salicylato)–
metal complexes, we recently reported the preparation
and crystal structures of (5-nitrosalicylato) complexes of
Mg(II), Zn(II), Co(II) and Ni(II). The building units of
these complexes self-assemble via p–p stacking and exten-
sive intermolecular hydrogen bonding, resulting in a 3D
supramolecular architecture [5]. Although the 5-nitrosali-
cylate ligand, 5-nsa, has a very versatile coordination
behavior, an exciting feature in this series is the possibility
of modifying the original structure by moving groups
attached to the aromatic nucleus. By shifting the hydroxyl
group from C-2 to the C-4 position of the 5-nsa ligand, a
Besides their relevance to solid state structure and
supramolecular design [1–3], (salicylato)–metal complexes
have attracted considerable attention, because of their
promising anti-inflammatory and anticonvulsant activities
*
Corresponding authors. Tel.: +33 146835699; fax: +33 146835752
(J. d’Angelo); tel.: +33 146835463 (G. Morgant); tel.: +1 501 686 6494;
fax: +1 501 526 6510 (J.R.J. Sorenson).
E-mail addresses: jean.dangelo@u-psud.fr (J. d’Angelo), georges.
morgant@u-psud.fr (G. Morgant), sorensonjohnrj@uams.edu (J.R.J.
Sorenson).
0277-5387/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2007.10.006

538
J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
suitable set of ligands for intermolecular anchoring
through hydrogen bonding was thus obtained for our
intended evaluations.
uncontrolled variation of pressure. Calibration of tempera-
ture was performed with 5 N quality indium at 156.6 ꢁC
and tin at 231.9 ꢁC and a heating rate of 10 ꢁC minꢀ1
.
Herein we report the crystal structures of (3-nitro-4-
hydroxybenzoato) complexes of Zn(II) and Co(II). A key
aspect of these structures is that the individual units self-
assemble via extensive hydrogen bonding, leading to an
original 3D continuum. Although, in the strict sense, the
present complexes are not salicylate derivatives but 4-OH
regioisomers, their anticonvulsant properties were deter-
mined to examine the effect of this structural change on
the structural and physical properties of these complexes
in relation to their pharmacological effects. For the sake
of comparison, the anticonvulsant activities of related tran-
sition metalloelement 5-nsa complexes were also deter-
mined [5]. Interestingly, promising pharmacological
responses were obtained for these complexes as anticonvul-
sants. Critical to the knowledge of the mode of action of
present anticonvulsant metal complexes was the identifica-
tion of the actual pharmacologically active entity that
crosses the blood–brain barrier. Toward this end, this
report presents: (1) the conducting properties of the metal
complexes in dilute aqueous solution and (2) water/1-octa-
nol extraction experiments which revealed an unexpected
equilibrium between these complexes and the correspond-
ing free ligands.
Impedance measurements of aqueous solutions were per-
formed using a VoltaLab 80 (Radiometer Analytical) with
a 2-pole cell with two platinum plates. The signal ampli-
tude was 100 mV. A continuous potential of 0 mV was
imposed. Solutions, preserved under nitrogen, were placed
in a double walled glass cell ensuring thermal stabilization.
Measurements were obtained at 20 ꢁC. Magnetic measure-
ments in the 2–300 K temperature range were carried out
with a MPMS5 SQUID susceptometer (Quantum Design
Inc.).
2.3. Synthesis of complexes
2.3.1. C14H16N2O14Zn (1)
Sodium 3-nitro-4-hydroxybenzoate was prepared by
neutralizing 3-nitro-4-hydroxybenzoic acid (0.915 g,
5 mmol) with 5 ml of 1 M NaOH in deionized water. To
this solution was added ZnSO4 Æ 7H2O (0.720 g, 2.5 mmol)
in 3 ml of water. After a few days 1 was filtered, crystallized
from water, and dried over anhydrous magnesium sulfate.
Yield: 0.939 g (75%). Anal. Calc. for C14H16N2O14Zn (1):
C, 33.52; H, 3.21; N, 5.58. Found: C, 33.43; H, 3.18; N,
5.56%. IR (neat, cmꢀ1): 3254 (s), 1621 (s), 1587 (s), 1526
(s), 1420 (s), 1379 (s), 1323 (s), 1264 (s), 1178 (s), 1153
(s), 1118 (s), 1074 (s), 942 (m), 916 (m), 858 (m), 825 (m),
789 (s), 764 (s), 684 (s), 638 (s).
2. Experimental
2.1. Materials
2.3.2. C14H16N2O14Co (2)
High purity 3-nitro-4-hydroxybenzoic acid, zinc sulfate
heptahydrate and cobalt sulfate heptahydrate were pur-
chased from Acros (USA) and used without further
purification.
Sodium 3-nitro-4-hydroxybenzoate was prepared by
neutralizing 3-nitro-4-hydroxybenzoic acid (0.915 g,
5 mmol) with 5 ml of 1 M NaOH in deionized water. To
this solution was added CoSO4 Æ 7H2O (0.702 g, 2.5 mmol)
in 3 ml of water. After a few days 2 was filtered, crystallized
from water, and dried over anhydrous magnesium sulfate.
Yield: 0.943 g (76%). Anal. Calc. for C14H16N2O14Co (2):
C, 33.95; H, 3.25; N, 5.66. Found: C, 34.05; H, 3.18; N,
5.64%. IR (neat, cmꢀ1): 3259 (m), 1621 (s), 1586 (s), 1545
(s), 1524 (s), 1418 (s), 1381 (s), 1323 (s), 1264 (s), 1179
(s), 1153 (s), 1117 (s), 1073 (m), 938 (m), 915 (m), 859
(m), 825 (m), 807 (s), 789 (s), 763 (s), 685 (s), 638 (s).
2.2. Physical measurements
Elemental analyses (C, H and N) were performed with a
Perkin–Elmer 2400 analyzer. Infrared spectra were
recorded with a Bruker Vector 22 spectrometer. Thermo-
gravimetric analyses (TGA) were performed with a TA
Instruments TGA Q500 apparatus. Calibrations were per-
formed at different temperatures using Curie magnetic
transitions for the recommended materials: alumel
(163 ꢁC) and nickel (354 ꢁC). Mass calibrations were per-
formed using a standard mass of 100 mg. The furnace
was calibrated between 100 and 900 ꢁC to verify thermo-
couple performance. All experiments were performed
2.4. X-ray diffraction measurements
Diffraction data were collected with a Bruker-SMART
three axis diffractometer equipped with a SMART 1000
CCD area detector using graphite monochromated Mo
under dry nitrogen, with a flow rate of 6 · 10ꢀ2 l minꢀ1
.
Ka X-radiation (wavelength k = 0.71073 A) at 100.0(1)
˚
Analysis of compounds following nitrogen purging was
conducted using a heating rate of 20 ꢁC minꢀ1 in order to
determine percent mass loss with optimal precision. Differ-
ential Scanning Calorimetry (DSC) experiments were per-
K. The low temperature was reached by an evaporated
liquid nitrogen stream over the crystal, provided by the
Oxford Cryosystem device. Data collection and processing
were performed using Bruker SMART programs [11] and
empirical absorption correction was applied using SADABS
computer program [11]. The structure was solved by direct
methods using SIR97 [12], and refined by full-matrix least-
formed with
a
TA Instruments Universal V4.2E
apparatus under nitrogen. Samples were introduced in alu-
minum pans and covered with holed caps in view to avoid

J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
539
Table 1
trodes of 6 Hz, 44 mA or lower, for 3 s is used to
produce psychomotor seizures, following instillation of a
drop of 0.9% saline prior to this application of electric
current. Abolition of the visual stereotypic behavior char-
acterized as minimal clonic automatistic behaviors. Subcu-
taneous (sc) pentylenetetrazol (Metrazol) seizure threshold
test: a dose of pentylenetetrazol, a potent CNS stimulant
(Central Nervous System), which produces seizures in
greater than 97% of treated mice, 85 mg/kg of body mass,
was administered sc as a 0.5% solution in saline at the pos-
terior neck midline. Animals were observed for 30 min.
Failure to observe even a threshold seizure, a single episode
of clonic spasms of at least 5 s duration, is defined as anti-
convulsant activity. Rotating rod test: mice or rats are
placed on a 1 inch diameter knurled plastic rod rotating
at six revolutions per min. Normal mice and rats can
remain on a rod rotating at this speed indefinitely. Neuro-
logic toxicity, either CNS stimulation or depression, is evi-
denced by a failure of the mouse or rat to remain on the
rotating rod for 1 min.
Crystal data and structure refinement parameters for C14H16N2O14
(1, M = Zn and for 2, M = Co)
M
Compound
1
2
Empirical formula
Molecular weight
Color
C14H16N2O14Zn
501.67
yellowish
triclinic
P1
6.209(5)
10.071(5)
14.729(5)
83.579(5)
79.084(5)
89.427(5)
898.6(9)
2
1.854
1.452
512
0.17 · 0.22 · 0.27
1.42–29.78
ꢀ8, 8/ꢀ13, 13/0,
19
C14H16N2O14Co
495.22
pink
triclinic
P1
Crystal system
Space group
ꢀ
ꢀ
˚
a (A)
6.197(5)
5.047(5)
14.671(5)
83.992(5)
99.698(5)
89.944(5)
449.5(6)
1
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
3
˚
V (A )
Z
Dcalc (g cmꢀ3
)
1.829
1.040
253
Absorption coefficient (mmꢀ1
F(000)
)
Crystal size (mm)
h Ranges (ꢁ)
h/k/l
0.18 · 0.22 · 0.25
2.83–29.54
ꢀ6, 6/ꢀ8, 7/0, 19
T (K)
100
6383
4266
100
3064
1834
3. Results and discussion
Number of reflections collected
Number of reflections
used P 3r(I)
3.1. Description of the C14H16N2O14Zn (1) and
C14H16N2O14Co (2) structures based upon X-ray
crystallography
Parameters
281
142
R [F2]
0.0243
0.0581
0.9011
ꢀ0.64; 0.64
0.0019
0.0346
0.0821
0.8934
ꢀ0.51; 0.58
0.0003
wR [F2]
Goodness-of-fit on F2ꢀ3
Complexes 1 and 2 can be represented by the general
formula (C7H4NO5)2M(H2O)4, where M = ZnII or CoII.
The canonical drawings of these complexes are represented
in Fig. 1 and the CAMERON molecular structures in Figs. 2
and 3, with the atom numbering system used in Table 2
which lists the most relevant bond distances and angles.
The formula units of complexes 1 and 2 can be repre-
sented by (C7H4NO5)2M(H2O)4 where M = Zn or Co.
For complex 2, the asymmetric unit consists of the half
of formula unit. In both complexes, the coordination octa-
hedron around the metal atom involves one carboxylato O-
atom from each 3-nitro-4-hydroxybenzoate ligand (acting
as a monodentate linker) and the O-atoms of the four
water molecules harnessed to the metal atom. In
(C7H4NO5)2Zn(H2O)4 (1), the octahedron is slightly dis-
torted. The cis angles around the Zn center vary between
85.97(7)ꢁ and 92.95(7)ꢁ while the trans angles vary from
175.70(5)ꢁ to 178.68(5)ꢁ. The Zn–O(26) and Zn–O(36) dis-
˚
Residual density (e A
Refine ls shift/su.max
)
squares based on F2 using CRYSTALS software [13]. The mol-
ecule was drawn using the CAMERON program [14]. All non-
hydrogen atoms were anisotropically refined. Hydrogen
atoms were located in difference Fourier maps. H atoms
were refined isotropically with Uiso = 1.20Ueq where Ueq
is the equivalent isotropic atomic displacement parameter
of the attached atom. Crystal parameters, data collection
and the refinement details of compounds 1 and 2 are
reported in Table 1.
2.5. Determination of anticonvulsant activities and Rotorod
toxicity
Anticonvulsant activities and Rotorod toxicity of
C14H16N2O14Zn (1) and C14H16N2O14Co (2), as well as
previously reported related complexes 3–6 [5], were deter-
mined by the National Institute of Health-National Insti-
tute of Neurological Disorders and Stroke-Antiepileptic
Drug Development program for the detection and evalua-
tion of compounds as anticonvulsant agents (USA). Max-
imal electroshock seizure were elicited with a 60 cycle
alternating current of 50 mA, five to seven times that nec-
essary to elicit minimal electroshock seizures, delivered
for 0.2 s via corneal electrodes. Psychomotor seizures test:
the delivery of a less intense electric stimulation, which pro-
duces less intense brain inflammation, via corneal elec-
˚
tances, 2.074(2) and 2.052(2) A, respectively, are notably
shorter than those for Zn–O(10), Zn–O(11), Zn–O(12),
and Zn–O(13); 2.186(2), 2.114(2), 2.092(2), and
˚
2.118(2) A, respectively.
O
OH
H2O
OH2
O2N
HO
O
M
O
NO2
H2O
OH2
O
Fig. 1. Canonical representation of (C7H4NO5)2M(H2O)4 (1, M = Zn; 2,
M = Co).

540
J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
H102
H112
O22
O10
O24
H21
N22
H101
H23
O11
H111
C23
O23
O37
C31
H35
O26
C35
C30
C22
H34
O33
C36
C21
Zn1
H132
C34
H33
C20
O27
C33
C24
O36
C26
H31
C25
H25
H24
H122
C32
O32
O13
H131
N32
O12
O34
H121
Fig. 2. Labeled CAMERON diagram of (C7H4NO5)2Zn(H2O)4 (1). Ellipsoids are drawn at a 50% probability level.
H34
C34
H35
O37
H111
C35
C36
O11
O34
O33
H121
H112
O36
C30
C33
O12
N32
H122
O32
H33
H31
C31
H33
C31
H31
C32
N32
Co1
O12
H122
O36
C32
C36
H112
O32
C30
C35
H121
C33
C34
O33
O34
O11
O37
H35
H111
H34
Fig. 3. Labeled CAMERON diagram of (C7H4NO5)2Co(H2O)4 (2). Ellipsoids are drawn at a 50% probability level.
Table 2
In contrast, the structure of (C7H4NO5)2Co(H2O)4 (2)
˚
Selected bond distances (A) and bond angles (ꢁ) for (C7H4NO5)2Zn(H2O)4
(1) and (C7H4NO5)2Co(H2O)4 (2)
consists of centrosymmetric monomers with the metal
atom occupying the inversion center at 0, 1/2, 1/2 of the tri-
clinic unit cell. This symmetry is revealed by two identical
1
2
˚
Bond lengths
Zn–O(10)
Zn–O(11)
Zn–O(12)
Zn–O(13)
Zn–O(26)
Zn–O(36)
Bond lengths
Co–O(11)
Co–O(12)
Co–O(36)
metal–carboxylato O-atom lengths of 2.070(2) A and two
2.186(2)
2.114(2)
2.092(2)
2.118(2)
2.074(2)
2.059(2)
2.127(2)
2.100(2)
2.070(2)
pairs of metal–O (water) distances of 2.100(2) and
˚
2.070(2) A. On the other hand, the cis angles around the
Co center consist of six centrosymmetrically-related pairs,
while identical values were observed for the three trans
angles, 180.00(8)ꢁ. Since complexes 1 and 2 exhibit very
close structural features, only details of the packing in 1
were arbitrarily chosen for presentation and discussion in
the present paper. The presence of four water molecules
around the metal center led to a very extensive and compli-
cated H-bonding network. A list of the most probable
atoms involved in H-bonds with their related angles is
given in Tables 3 and 4. Fig. 4 shows the H-bonding in 1
around the Zn center. The crystal cohesion is thus ensured
by fourteen hydrogen bonds per formula unit.
Bond angles
Bond angles
O(10)–Zn–O(12)
O(11)–Zn–O(13)
O(26)–Zn–O(36)
O(10)–Zn–O(11)
O(10)–Zn–O(13)
O(10)–Zn–O(26)
O(10)–Zn–O(36)
O(12)–Zn–O(11)
O(12)–Zn–O(13)
O(12)–Zn–O(26)
O(12)–Zn–O(36)
O(11)–Zn–O(26)
O(11)–Zn–O(36)
O(13)–Zn–O(26)
O(13)–Zn–O(36)
177.58(6)
178.68(5)
175.70(5)
86.57(6)
92.61(6)
89.03(6)
86.98(6)
91.03(7)
89.78(6)
91.21(6)
92.84(6)
88.42(7)
92.95(7)
92.60(7)
85.97(7)
O(11)vi–Co–O(11)
O(12)vi–Co–O(12)
O(36)vi–Co–O(36)
O(11)vi–Co–O(36)vi
O(11)–Co–O(36)
O(11)vi–Co–O(12)vi
O(11)–Co–O(12)
O(12)vi–Co–O(11)
O(11)vi–Co–O(12)
O(36)vi–Co–O(12)
O(12)vi–Co–O(36)
O(11)vi–Co–O(36)
O(36)vi–Co–O(11)
O(12)vi–Co–O(36)vi
O(12)–Co–O(36)
180.00(8)
180.00(8)
180.00(8)
93.26(8)
93.26(8)
88.40(9)
88.40(9)
91.60(9)
91.60(9)
89.68(9)
89.68(9)
86.74(8)
86.74(8)
90.32(9)
90.32(9)
The individual zero-dimensional units (0D) of 1 are
linked together through intermolecular hydrogen bonding
implying their ortho-nitrophenol groups ends, O(23)iv–
H(23)ivꢁ ꢁ ꢁO(32) and O(33)–H(33)ꢁ ꢁ ꢁO(22)iv forming infi-
nite 1D chains (Figs. 5 and 6). The intrachain separation
vi
˚
˚
Symmetry code: ꢀx, ꢀy + 1, ꢀz + 1.
of the Zn centers is 18.82 A (18.85 A for Co–Co in complex
2). It is noteworthy that the ortho-nitrophenol groups are
also implied in intramolecular hydrogen bonding,
O(23)iv–H(23)ivꢁ ꢁ ꢁO(22)iv and O(33)–H(33)ꢁ ꢁ ꢁO(32), which
clearly reinforces the cohesion of these chains.
The parallel chains are further engaged in hydrogen
bonding implying H(102), H(112) and H(121) with inter-
chain separations of 4.904 A and 5.200 A (5.047 A for com-
˚
˚
˚

J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
541
Table 3
H-bond donor/acceptor scheme (A, ꢁ) for (C7H4NO5)2Zn(H2O)4 (1)
complex 2) yielding a 3D supramolecular architecture
(Fig. 8).
˚
D–Hꢁ ꢁ ꢁA
d(D–H)
D(Hꢁ ꢁ ꢁA)
\(DHA)
One of the most salient features of this architecture is
the compact 3D arrangement of the metal atoms. In com-
plex 1, the Zn atoms are close to each other, with distances
O(23)–H(23)ꢁ ꢁ ꢁO(22)
O(33)–H(33)ꢁ ꢁ ꢁO(32)
O(11)–H(111)ꢁ ꢁ ꢁO(37)
O(13)–H(131)ꢁ ꢁ ꢁO(27)
O(10)–H(101)ꢁ ꢁ ꢁO(37)i
O(13)–H(132)ꢁ ꢁ ꢁO(37)i
O(12)–H(122)ꢁ ꢁ ꢁO(27)ii
O(11)–H(112)ꢁ ꢁ ꢁO(26)iii
O(10)–H(102)ꢁ ꢁ ꢁO(11)iii
O(33)–H(33)ꢁ ꢁ ꢁO(22)iv
O(23)iv–H(23)ivꢁ ꢁ ꢁO(32)
O(23)–H(23)ꢁ ꢁ ꢁO(32)v
O(33)v–H(33)vꢁ ꢁ ꢁO(22)
O(12)–H(121)ꢁ ꢁ ꢁO(13)vi
i
0.795(2)
0.815(2)
0.940(2)
0.940(2)
0.934(2)
0.940(2)
0.932(2)
0.939(2)
0.936(2)
0.814(2)
0.794(2)
0.794(2)
0.814(2)
0.927(2)
1.865(2)
1.902(2)
1.714(2)
1.720(2)
1.939(2)
1.769(2)
1.912(2)
1.779(2)
1.938(2)
2.345(2)
2.401(2)
2.401(2)
2.345(2)
1.888(2)
153.13(9)
143.30(9)
161.11(9)
160.63(8)
147.17(9)
164.39(9)
166.64(9)
166.97(8)
154.65(8)
139.78(9)
130.98(9)
130.98(9)
139.78(9)
174.50(8)
iv
˚
of 4.904, 5.200 and 6.209 A (Fig. 9a). A centrosymmetric
metal arrangement pattern was observed for complex 2
˚
with Co – Co distances of 5.047 and 6.197 A (Fig. 9b).
3.2. Thermal analysis: TGA and DSC measurements
TGA analysis of 1 revealed a mass loss of 14.8% on
heating from 50 to 180 ꢁC, corresponding to a loss of four
crystalline water molecules (Calc. 14.4%). TGA analysis of
2 revealed a mass loss of 15.1% on heating from 80 to
180 ꢁC, corresponding to a loss of four crystalline water
molecules (Calc. 14.5%). On heating 2 from 80 to 550 ꢁC,
an overall mass loss of 85.5% was observed. The non-vol-
atile residue, 14.5% of the starting material, corresponds
to one molecule of CoO (Calc. 15.1%). The DSC study of
1 revealed, as the main event, a narrow endotherm centered
at 141 ꢁC corresponding to loss of water and a narrow exo-
therm centered at 342 ꢁC. A similar thermal pattern was
observed for complex 2, with an endotherm centered at
130 ꢁC and exotherm centered at 349 ꢁC.
ii
iii
Symmetry codes: x + 1, y, z; x ꢀ 1, y + z; ꢀx, ꢀy, ꢀz + 1; x ꢀ 1,
v
vi
y + 1, z + 1; x + 1, y ꢀ 1, z ꢀ 1; ꢀx, ꢀy + 1, ꢀz + 1.
Table 4
˚
H-bond donor/acceptor scheme (A, ꢁ) for (C7H4NO5)2Co(H2O)4 (2)
D–Hꢁ ꢁ ꢁA
d(D–H)
D(Hꢁ ꢁ ꢁA)
\(DHA)
O(11)–H(111)ꢁ ꢁ ꢁO(37)
O(11)–H(112)ꢁ ꢁ ꢁO(36)i
O(12)–H(122)ꢁ ꢁ ꢁO(11)ii
O(12)–H(121)ꢁ ꢁ ꢁO(37)iii
O(33)–H(33)ꢁ ꢁ ꢁO(32)
O(33)–H(33)ꢁ ꢁ ꢁO(32)iv
O(33)iv–H(33)ivꢁ ꢁ ꢁO(32)
i
0.954(4)
0.951(7)
0.944(7)
0.947(5)
0.810(5)
0.810(5)
0.810(5)
ii
1.673(2)
2.162(3)
1.860(3)
1.834(3)
1.898(3)
2.346(2)
2.346(2)
163.01(8)
158.83(6)
170.02(9)
161.37(9)
144.75(6)
138.38(5)
138.38(5)
3.3. Electrochemical impedance measurements
iii
Symmetry codes: x + 1, y, z; x, y ꢀ 1, z; ꢀx + 1, ꢀy + 1, ꢀz + 1;
The electrochemical impedance Z(x) of the studied sam-
ple between the two plates of the conductivity cell is a com-
plex number which can be represented, in polar coordinates
by its module |Z| and its phase, in Cartesian representations
by its real part Z0 and its imaginary part Z00 with
Z(x) = Z0 + i Z00. At the time of frequency scanning, the
experimental points were distributed on a curve having
the form of an arc of a circle. The obtained curve is called
a Nyquist diagram (ꢀZ00 versus Z0). Z01 to high frequency,
HF, and Z02 to low frequency, LF, are the impedance when
‘Z’ crosses the OZ axis (real axis). The Nyquist curves for
1 · 10ꢀ4 M solutions of complexes 1 and 2 (a concentration
close to the doses administered in the anticonvulsant and
toxicity assays, see Section 3.6) are shown in Fig. 10. The
existence of a half circle is typical of the presence of a sim-
ple charge transfer. Table 5 gives the electrochemical con-
ductances (1/Z0) of complexes 1 and 2 at concentrations
of 1 · 10ꢀ4 M and 5 · 10ꢀ4 M in deionized water. At LF,
conductances ð1=Z02Þ of solutions of complexes 1 and 2 at
identical concentrations are very similar (compare entries
1 and 5 in one hand, and 3 and 7 in the other hand). These
values are notably higher than the conductance of deion-
ized water employed in the present experiments (entry 9),
that is to say, approximately 25 times for a concentration
of 1 · 10ꢀ4 M and 130 times for a concentration of
5 · 10ꢀ4 M. Note that at LF, the cell capacitance is
assumed to be so small that it may be neglected. However,
at HF, conductances ð1=Z01Þ for solutions of both com-
plexes are of the same order of magnitude, irrespective of
iv
ꢀx ꢀ 2, ꢀy + 2, ꢀz + 2.
Fig. 4. Labeled diagram of (C7H4NO5)2Zn(H2O)4 (1) focusing on the
hydrogen bonding system around the Zn atom (dashed lines).
plex 2) resulting in 2D sheets (Fig. 7). These, in turn, self-
assemble via hydrogen bonds implying H(122) and H(132)
with an interplanar separation of 6.209 A (7.996 A for
˚
˚

542
J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
1D
Fig. 5. Part of [(C7H4NO5)2Zn(H2O)4]n (1)n showing the 1D chain structure with intrachain hydrogen bonds (dashed lines). Zn centers are symbolized by
large circles.
concentrations (entries 11 and 12). Complexes 1 and 2
are therefore strong electrolytes.
O34
N32
C24
O23
C31
O32
3.4. Magnetic measurements
C23
iv
H23
C32
The magnetic properties of Co-complex 2 are shown in
Fig. 11. At room temperature, vMT is equal to
3.02 cm3 K molꢀ1, value expected for an isolated Co(II)
ion. Upon cooling, the vMT values steadily decrease up
to 7 K. Below this temperature, the decrease of vMT is stee-
per to reach a value of 1.62 cm3 K molꢀ1 at 2 K. In fact, in
the 300–30 K temperature range, the experimental curve
can be fitted using the theoretical behavior of a distorted
octahedral isolated cobalt ion [15]. The best fit of the data
by full-matrix diagonalization of the appropriate spin
Hamiltonian is obtained with the following parameter
k = ꢀ117, D = 605 and j = 1 where k, D and j stand for
spin–orbit constant, axial distortion and orbital reduction
factor, respectively. As shown in the inset of Fig. 11, below
30 K the experimental data deviate from the theoretical
curve. First, between 30 and 3 K the experimental data
are above the theoretical curve indicating that ferromag-
netic interaction between Co ions is operative within the
C22
iv
O22
C33
H33
C21
N22
O33
C34
O24
Fig. 6. Labeled diagram of (C7H4NO5)2Zn(H2O)4 (1) focusing on the
hydrogen bonding system around the ortho-nitrophenol groups (dashed
lines).
concentrations of these complexes (entries 2, 4, 6 and 8),
and close to that of deionized water (entry 10). Clearly,
at HF, the cell capacitance now markedly competes with
the cell resistance, leading to fatal errors in conductance
measurements. LF measurements indicate that the conduc-
tances of the 1 · 10ꢀ4 M and 5 · 10ꢀ4 M complex solutions
do not differ greatly from those of KCl solutions at these
1D
2D
Fig. 7. Packing of (C7H4NO5)2Zn(H2O)4 (1) showing the H-bonded 2D continuum as dashed lines. Zn centers are symbolized by large circles.

J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
543
3D
1D
2D
Fig. 8. Packing of (C7H4NO5)2Zn(H2O)4 (1) showing the H-bonded 3D continuum as dashed lines. Zn centers are symbolized by large circles.
Fig. 9. Packing of (C7H4NO5)2Zn(H2O)4 (1) (a) and (C7H4NO5)2Co(H2O)4 (2) (b) concentrating on the 3D arrangement of the metal atoms. Hydrogen
iii
vi
i
ii
˚
˚
˚
bonds are symbolized by dashed lines. Metal–metal distances for Zn–Zn : 4.904 A, Zn–Zn , 5.200 A and Zn–Zn = Zn–Zn , 6.209 A. Symmetry codes:
iii
ii
˚
˚
i: x + 1, y, z; ii: x ꢀ 1, y + z; iii: ꢀx, ꢀy, ꢀz + 1; vi: ꢀx, ꢀy + 1, ꢀz + 1. Metal–metal distances for Co–Co : 5.047 A and Co–Co , 6.197 A. Symmetry
codes: iii: ꢀx + 1, ꢀy + 1, ꢀz + 1; ii: x, y ꢀ 1, z.
crystal. Secondly, at very low temperature (3–2 K) the
experimental values becomes lower than the theoretical
curve showing that a weak antiferromagnetic interaction
is also present in 2. These results are in agreement with
the two first levels of hydrogen bonds network in Co-com-
plex 2 (Fig. 9b). On this basis, one can conclude that, mag-
netically, the cobalt ions are ferromagnetically coupled
through the hydrogen bonds network of water molecules
obtained with Co-complex 2. Complex 1 (306 mg,
6.1 · 10ꢀ4 mol) was dissolved in deionized water (15 ml).
iii
˚
(Co –O11–O10–Co, Co–Co = 5.047 A) forming 1D ferro-
magnetic chains which are antiferromagnetically coupled
by the hydrogen bonds developed through the ortho-nitro-
phenol groups of 3-nitro-4-hydroxybenzoate ligand.
3.5. Water/1-octanol extraction experiments
Only the details of the extraction experiment involving
Zn-complex 1 are given below. Similar results were
Fig. 10. Nyquist diagrams of complexes 1 and 2 at concentration of
1 · 10ꢀ4 M.

544
J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
Table 5
color of the aqueous phase changed from pale yellow to
bright orange. Evaporation of this solution at room tem-
perature under vacuo gave a microcrystalline orange solid
which was successively washed with THF and diethyl ether.
Elemental analysis of this solid (Found: C, 30.71; H, 2.41;
N, 5.05; Zn, 22.99%) is consistent with the molecular for-
mula C7H3NO5Zn(H2O)2 (Calc: C, 29.79; H, 2.48; N,
4.96; Zn, 23.14%) corresponding to a salt of type D
(Fig. 12) in which the ligand C and Zn atom components
are in the ratio of 1 to 1 (while starting complex 1 com-
prises two ligands C per Zn atom). Two water molecules
are linked to each Zn atom of this salt, in order to satisfy
the tetrahedral coordination of the metal. The IR spectrum
of this salt, (neat, cmꢀ1): 3226 (m, m O–H), 1596 (s, mas
COOꢀ), 1518 (s), 1391 (s, ms COOꢀ), 1324 (s), 1248 (s),
1192 (m), 1153 (s), 1073 (m), 927 (m), 831 (m), 783 (s),
763 (s), 709 (s), 643 (s), is in agreement with the proposed
molecular structure. The above outcome can be interpreted
on the basis of the putative two-step equilibrium of com-
plexes 1 and 2 in aqueous medium depicted in Fig. 12.
The first step of this process involves the dissociation of
the complexes into cation A and anion B [16] (note that
similar discrete ionic entities were found in the crystal
structures of 5(nsa)-metal complexes 3 and 4, regioisomers
of 1 and 2, respectively, see Fig. 13). These, in turn, may
interconvert through proton transfer into ligand C and salt
D (in which each metal atom is coordinated to two ligand
molecules acting as bidendate linkers through their car-
boxylato and phenoxo groups). Because of its marked lipo-
Electrochemical conductance for complexes 1 and 2 at 20 ꢁCa
Entry
Compound
Concentration
(M)
Frequency
(kHz)
1/Z0
(Xꢀ1 cmꢀ1
)
1
2
3
4
5
6
7
8
9
1
1 · 10ꢀ4
1 · 10ꢀ4
5 · 10ꢀ4
5 · 10ꢀ4
1 · 10ꢀ4
1 · 10ꢀ4
5 · 10ꢀ4
5 · 10ꢀ4
3
>100
18
>100
3
>100
28
>100
0.01
18
2.02 · 10ꢀ5
3.57 · 10ꢀ4
9.90 · 10ꢀ5
4.54 · 10ꢀ4
2.16 · 10ꢀ5
3.33 · 10ꢀ4
1.18 · 10ꢀ4
3.57 · 10ꢀ4
8.35 · 10ꢀ7
2.22 · 10ꢀ4
1.47 · 10ꢀ5
7.98 · 10ꢀ5
1
1
1
2
2
2
2
Deionized water
Deionized water
KCl
10
11
12
a
1 · 10ꢀ4
5 · 10ꢀ4
1.4
0.12
KCl
Cell constant = 0.973.
philicity (calculated logP = 1.78) [17], ligand
C is
extractable with 1-octanol (an operation that drives the
equilibrium towards C and D partners), while hydrophilic
salt D remains in the aqueous phase.
Fig. 11. vMT vs. T plot of Co-complex 2, experimental (e) and calculated
(——). The inset shows the deviation of the experimental vMT values from
the Co monomer behavior.
This solution was extracted with 1-octanol (15 ml) at room
temperature. The extraction was repeated four times and
the 1-octanol extracts combined. After removal of 1-octa-
nol by vacuum distillation at 1 Torr, 3-nitro-4-hydroxyben-
zoic acid (111 mg, 6.1 · 10ꢀ4 mol), compound C in Fig. 12,
was obtained as a pale yellow solid, which was unambigu-
ously identified by comparison with an authentic sample
using Rf values in thin layer chromatography and IR,
OH2
H2O
H2O
OH2
O
M
NO2
H2O
NO2
. H2O
HO
O
O
O
HO
Fig. 13. Crystal structure-based canonical representation of [M(H2O)5-
(5-nsa)]+(5-nsa)ꢀ Æ H2O (3, M = Zn; 4, M = Co; 5, M = Ni; 6, M = Mg) [5].
1
UV, and H NMR spectra. During this experiment, the
O2C
OH
NO2
HO
O2N
CO2M
+
1
or
2
B
A
HO
CO2H
-M-O
O2N
CO2 -M-O
CO2-
+
O2N
O2N
C
D
Fig. 12. Putative equilibrium of complexes 1 and 2 in aqueous solution (M = Zn or Co). The water molecules coordinated to the metal atoms have been
omitted for clarity.

J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
545
Table 6
Comparative anticonvulsant activities and Rotorod toxicity for complexes 1 to 6 (lmol/kg of body mass)
Compound [Reference]
MESa
Minimal clonic seizureb
scMETc
Rotorod toxicityd
1 [this work]
2 [this work]
I at 600 ipm at 0.5 h
I at 610 ipm at 0.5 h
I at 560 ipm at 0.5 h
A at 93 or at 0.25 h
I at 570 ipm at 4 h
I at 600 ipm at 4 h
A at 198 ipm at 0.25 h
A at 200 ipm at 1 h
A at 185 at 0.25 h
A at 185 ipm at 2 h
A at 57 ipm at 1 h
A at 199 ipm at 0.5 h
A at 198 ipm at 0.25 h
A at 200 ipm at 0.5 h
A at 185 ipm at 1 h
I at 560 ipm at 4 h
I at 570 ipm at 4 h
I at 600 ipm at 4 h
A at 700 ipm at 1 h
A at 610 at 0.5 h
A at 185 ipm at 0.25 h
A at 185 ipm at 0.5 h
A at 57 ipm at 0.5 h
I at 600 ipm at 4 h
3 [5]
4 [5]
5 [5]
6 [5]
a
MES = maximal electroshock seizure test.
b
Psychomotor seizure or minimal clonic seizures are usually produced with the delivery of 6 Hz (32 mA) or slightly higher current designed to detect
compounds that are missed with the MES tests which employs a larger current of 50 mA. It is a much less intense stimulation of the CNS than the MES
test.
c
scMET = subcutaneous Metrazol seizure threshold test, and Rotorod toxicity test.
Rotorod toxicity, an inability to grasp the rotating rod, may be due to CNS depression due to sedative and/or hypnotic activities, or a state of eminent
d
death, due to medullary paralysis at high doses, a common unwanted side effect caused by high doses of all anticonvulsant drugs. A = Activity at indicated
dose and time of challenge. I = Inactive, ip = intraperitoneal, o = oral, m = mice, r = rats.
3.6. Anticonvulsant activities
1-octanol and water [18]. Because of their marked hydro-
philicity, revealed by their notable water solubility (Section
2.3) and the important conductance of their aqueous solu-
tions (Section 3.3), metal salts 1 and 2 can have direct
access to the systemic circulation via blood plasma. Lipo-
philic ligand C, which is in equilibrium with complexes 1
and 2 in aqueous medium (Section 3.5), can then penetrate
the BBB by diffusion (passive translocation via a lipophilic
pathway). On this basis, we may hypothesize that the active
form of these anticonvulsant metal complexes might be free
ligand C, the metal salts playing the role of hydrosoluble
vectors that release the drug at BBB level. This hypothesis
is in full agreement with a previous investigation of (salicy-
lato)–copper complex equilibrium under physiological con-
ditions, which ruled out any implication of the copper
complex in the anti-inflammatory activity, since the salicy-
late anion cannot significantly mobilize plasma copper into
tissue–diffusible complexes. Even though the salicylate
ligand can effectively promote the gastrointestinal absorp-
tion of copper, the reciprocal effect of the metal upon the
drug seems to be negligible, except for high metal-to-drug
ratios [19–21].
Anticonvulsant activities for compounds 1 and 2, as well
as related (5-nitrosalicylato)–metal complexes [5] (com-
pounds 3–6, Fig. 13), are presented in Table 6.
Only 4 was found to have activity against MES-induced
seizures. This modest activity was observed at a relatively
low and non-toxic dose given orally to rats. Compounds
1–6 had activity in protecting against the less intense min-
imal clonic seizure with an apparent activity order being
5 > 2 @ 1 > 3 @ 6 and 4, based upon these preliminary
results. Only compounds 1, 2and 3 afforded some protec-
tion against the MET-induced seizure and the relative
order of effectiveness was 1 > 2 > 3.
4. Conclusion
We have shown that the individual subunits of the (3-
nitro-4-hydroxybenzoato) complexes of Zn(II) (1) and
Co(II) (2) self-assemble via extensive hydrogen bonding
involving water molecules leading to a 3D supramolecular
architecture. The electrochemical impedance measurements
show that these complexes are strong electrolytes in dilute
aqueous solution. On the other hand, magnetic measure-
ment of Co-complex indicates that the Co ions are ferro-
magnetically coupled thought the hydrogen bonds
network of water molecules forming 1D ferromagnetic
chains.
The anticonvulsant activities of complexes 1 and 2 as
well as related Zn(II), Co(II), Ni(II) and Mg(II) (5-nitrosal-
icylato) complexes (3, 4, 5 and 6, respectively) where also
determined. The pharmacological responses obtained for
these complexes are consistent with those obtained for
many related metal complexes [4–10].
Acknowledgements
´
Financial support of Universite Paris XI, CNRS and
Ecole Centrale Paris is gratefully acknowledged. The
authors thank Mrs. S. Mairesse-Lebrun for performing ele-
mental analyses. We are also thank Mr. James Stables for
his assistance in obtaining data pertain to the anticonvul-
sant and Rotorod toxicities of our compounds via the
DHHS, NIH, NINDS ADD program (USA).
Appendix A. Supplementary material
The transport of therapeutics across the blood–brain
barrier (BBB) is critical for the treatment of disorders of
the central nervous system. There is a good correlation
between BBB penetration in vivo and lipid solubility of a
drug, the latter being conventionally expressed as its parti-
tion coefficient (P), calculated as the ratio of solubilities in
CCDC 647852 and 647853 contain the supplemen-
tary crystallographic data for 1 and 2. These data can be
obtained free of charge via http://www.ccdc.cam.ac.uk/
conts/retrieving.html, or from the Cambridge Crystallo-
graphic Data Centre, 12 Union Road, Cambridge CB2

546
J. d’Angelo et al. / Polyhedron 27 (2008) 537–546
[10] B. Viossat, F.T. Greenaway, G. Morgant, J.-C. Daran, N.-H. Dung,
J.R.J. Sorenson, J. Inorg. Biochem. 99 (2005) 355.
[11] ASTRO (5.00), SAINT (5.007) and SADABS (5.007). Data Collection
and Processing Software for the SMART System (5.054). Siemens
(BRUKER-AXS) Analytical X-ray Instruments Inc., Madison,
WI, 1998.
1EZ, UK; fax: (+44) 1223-336-033; or e-mail: deposit@
ccdc.cam.ac.uk. Supplementary data associated with this
article can be found, in the online version, at
doi:10.1016/j.poly.2007.10.006.
[12] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, G. Giaco-
vazzo, A. Guagliardi, A. Grazia, G. Moliterni, G. Polidori, R.
Spagna, J. Appl. Crystallogr. 32 (1999) 115.
References
[1] F.D. Rochon, G. Massarweh, Inorg. Chim. Acta 304 (2000) 190.
[2] S.H. Rahaman, D. Bose, H. Chowdhury, G. Mostafa, H.-K. Fun,
B.K. Ghosh, Polyhedron 24 (2005) 1837.
[13] D.J. Watkin, C.K Prout, J.R. Carruthers, P.W. Betteridge, R.I.
Prout, CRYSTALS Issue 11. Chemical Crystallography Laboratory,
University of Oxford, UK, 2001.
[3] N.E. Ghermani, G. Morgant, J. d’Angelo, D. Desmae¨le, B. Fraisse,
F. Bonhomme, E. Dichi, M. Sghaier, Polyhedron 26 (2007) 2880.
[4] J.R. J Sorenson, Free Radic. Biol. Med. 13 (1992) 593.
[5] G. Morgant, N. Bouhmaida, L. Balde, N.E. Ghermani, J. d’Angelo,
Polyhedron 25 (2006) 2229.
[14] D.J. Watkin, C.K. Prout, L.J. Pearce, CAMERON, Chemical Crystal-
lography Laboratory, University of Oxford, UK, 1996.
[15] O. Kahn, Molecular Magnetism, VCH, New York, 1993.
[16] L.H.J. Lajunen, R. Portanova, J. Piispanen, M. Tolazzi, Pure Appl.
Chem. 69 (1997) 329.
[6] J.R.J. Sorenson, J. Med. Chem. 19 (1976) 135.
[17] N. Bodor, P. Buchwald, J. Phys. Chem. B 101 (1997) 3404.
[18] N.J. Abbott, I.A. Romero, Mol. Med. Today (1996) 106.
[19] V. Brumas, B. Brumas, G. Berthon, J. Inorg. Biochem. 57 (1995)
191.
[20] H. Miche, V. Brumas, G. Berthon, J. Inorg. Biochem. 68 (1997) 27.
[21] B. Halova-Lajoie, V. Brumas, M.M.L. Fiallo, G. Berthon, J. Inorg.
Biochem. 100 (2006) 362.
[7] J.R.J. Sorenson, Prog. Med. Chem. 15 (1978) 211.
[8] B. Viossat, J.-C. Daran, G. Savouret, G. Morgant, F.T. Greenaway,
N.-H. Dung, V.A. Pham-Tran, J.R.J. Sorenson, J. Inorg. Biochem. 96
(2003) 375.
[9] P. Lemoine, B. Viossat, N.H. Dung, A. Tomas, G. Morgant, F.T.
Greenaway, J.R.J. Sorenson, J. Inorg. Biochem. 98 (2004) 1734.