Powder diffraction study of the hydrogen bonds in nitroxoline and its hydrochloride.

Article abstract of DOI:10.1107/S0108270101018224

The crystal structures of 8-hydroxy-5-nitroquinoline, C9H6N2O3, (I), and 8-hydroxy-5-nitroquinolinium chloride, C9H7N2O3+*Cl-, (II), have been determined from X-ray powder data. In (I), the molecules are linked via moderately strong hydrogen bonds to form dimers. Such a packing motif is likely to be responsible for the low solubility of (I) in water. In (II), the inversion-related cations form stacks, and anions fill the interstack channels.

Full text of DOI:10.1107/S0108270101018224

organic compounds
Acta Crystallographica Section C
Crystal Structure
Communications
absorb light in essentially the whole of the visible region. In
practice, crystals of (I) are light yellow, with no absorption
maximum in the visible region, and the hydrochloride salt,
(II), is deeper in colour than (I), in full agreement with the
results of the calculations for the neutral form and cation, and
with the spectroscopic data for solutions (Ermakov et al.,
1985).
ISSN 0108-2701
Powder diffraction study of the
hydrogen bonds in nitroxoline and
its hydrochloride
a
a
A. V. Yatsenko, * K. A. Paseshnichenko, V. V.
a
b
Chernyshev and H. Schenk
a
Chemistry Department, Moscow State University, 119899 Moscow, Russia, and
The molecule of (I) is nearly planar, the dihedral angle
formed by the least-squares planes through the naphthalene
b
Laboratory for Crystallography, University of Amsterdam, Nieuwe Achtergracht 166,
Amsterdam 1018 WV, The Netherlands
ꢀ
and nitro moieties being only 6.3 (4) . The dihedral angle
formed by the planes through the naphthalene moieties of
Correspondence e-mail: yatsenko@biocryst.phys.msu.su
ꢀ
Received 21 September 2001
Accepted 26 October 2001
Online 14 December 2001
adjacent molecules is 44.8 (5) . Details of the hydrogen bonds
are given in Table 1. Molecules related by the 2 axis form
dimers, as shown in Fig. 1. We have carried out the DFT
optimization of one such dimer, starting from the crystal-
lographic geometry. The resulting geometry does not signi®-
cantly differ from the experimental geometry. After the
optimization, the angle subtended by two naphthalene
The crystal structures of 8-hydroxy-5-nitroquinoline, C H -
9
6
N O , (I), and 8-hydroxy-5-nitroquinolinium chloride, C H -
2
3
9
7
+
�
N O ÁCl , (II), have been determined from X-ray powder
2
3
data. In (I), the molecules are linked via moderately strong
hydrogen bonds to form dimers. Such a packing motif is likely
to be responsible for the low solubility of (I) in water. In (II),
the inversion-related cations form stacks, and anions ®ll the
interstack channels.
ꢀ
moieties is increased to 64 and the hydroxy group is twisted
ꢀ
by 15 with respect to the naphthalene plane, thus facilitating
the hydrogen bonding: the O8ÐH8Á Á ÁN1 angle is increased to
ꢀ
ꢀ
ꢀ
1
55 , the C9ÐN1Á Á ÁH8 angle becomes 129 (versus 135 in the
Comment
Nitroxoline, or 8-hydroxy-5-nitroquinoline, (I), also known as
NOK, shows antibacterial and fungicidal activity. In parti-
5
cular, it is used for the treatment of infections of the genito-
urinary system. Compound (I) has low solubility in water, but
its solubility increases considerably in alkalis and acids.
To date, it has not been clear whether the molecule of (I) in
the crystal exists in the neutral form, or as a zwitterion
resulting from proton transfer from the hydroxy group to the
quinoline N atom. DFT (density functional theory) calcula-
tions for isolated molecules predict that the zwitterionic form
is 31 kJ mol less stable than the neutral form. This differ-
ence, while signi®cant, is not large enough to exclude the
zwitterion from consideration, since it can be overshadowed
by the effects of the medium. However, visible spectroscopy,
together with semi-empirical calculations on the INDO±CISD
level (Dick & Nickel, 1983), give strong evidence for the
presence of the neutral molecule. Recently, the INDO±CISD
technique was proven to be suitable for reproducing the
spectra of nitronaphthalenoles (Yatsenko et al., 2001). For (I),
INDO±CISD calculations predict that the neutral form should
have its ꢀ±ꢀ absorption maximum in the near-UV region, at
Figure 1
The hydrogen-bonded dimers in (I) showing the atom-labelling scheme.
�
1
3
07 nm. Protonation of the quinoline N atom shifts this band
bathochromically by 40 nm, whereas in the case of the zwit-
terion, this maximum is shifted to 391 nm and a new absorp-
tion band appears at 635 nm. Thus, the zwitterion should
Figure 2
The ClÁ Á ÁH contacts in (II); the atom-labelling scheme is the same as for
(I) in Fig. 1.
Acta Cryst. (2002). C58, o19±o21
DOI: 10.1107/S0108270101018224
# 2002 International Union of Crystallography o19

organic compounds
giving very thin needles. The sample obtained from aqueous solution
gave an identical diffraction pattern, although of lower quality.
Compound (II) was prepared by recrystallization of (I) from 5%
HCl.
Compound (I)
Crystal data
C
9
H
6
N
2
O
3
Cell parameters from 26
re¯ections
r
M = 190.16
ꢀ
Orthorhombic, Fdd2
Ê
ꢂ = 4.3±16.4
ꢃ = 1.02 mm
T = 295 (2) K
� 1
a = 28.049 (7) A
Ê
b = 31.198 (8) A
Ê
c = 3.727 (1) A
Specimen shape: ¯at sheet
25 Â 25 Â 2.0 mm
Particle morphology: needle, light
yellow
Ê
V = 3261.4 (15) A
3
Z = 16
D = 1.549 Mg m
�
3
x
Figure 3
Cu Kꢁ radiation
The Rietveld plot for (I) showing the observed and difference pro®les.
The re¯ection positions are shown above the difference pro®le. The high-
angle area is magni®ed by a factor of 10.
Data collection
DRON-3M diffractometer (Bure-
vestnik, Russia)
Specimen mounting: pressed as a
thin layer in the specimen holder
Specimen mounted in re¯ection
mode
h = 0 ! 20
k = 0 ! 22
l = 0 ! 2
ꢀ
2ꢂmin = 7.0, 2ꢂmax = 70.0
ꢀ
Increment in 2ꢂ = 0.02
Re®nement
R
R
R
p
= 0.035
wp = 0.048
exp = 0.023
Pro®le function: split-type pseudo-
Voigt
224 re¯ections
S = 2.10
92 parameters
H-atom parameters constrained
(Á/ꢄ)max = 0.045
Preferred orientation correction:
spherical harmonics (Ahtee et al.,
1989)
ꢀ
2ꢂ
min = 7.90, 2ꢂmax = 70.0
ꢀ
Increment in 2ꢂ = 0.02
Wavelength of incident radiation:
Ê
.5418 A
1
Excluded region(s): 7.00±7.88
Table 1
Ê
ꢀ
Geometry of hydrogen bonds and short CÐHÁ Á ÁO contacts (A, ) for (I).
Figure 4
The Rietveld plot for (II) showing the observed and difference pro®les.
The re¯ection positions are shown above the difference pro®le. The high-
angle area is magni®ed by a factor of 10.
DÐHÁ Á ÁA
DÐH
HÁ Á ÁA
DÁ Á ÁA
DÐHÁ Á ÁA
i
O8ÐH8Á Á ÁN1
0.90
1.00
1.00
2.05
2.42
2.59
2.811 (5)
3.304 (7)
3.472 (6)
141
147
147
ii
C6ÐH6Á Á ÁO2
iii
Ê
C3ÐH3Á Á ÁO2
crystal) and the OÁ Á ÁN distance is 0.07 A shorter than in the
1
2
1
2
1
3
3
crystal. The DFT-calculated energy of the dimer is
1
Symmetry codes: (i) � x; 1 � y; z; (ii) � x; 1 � y; ‡ z; (iii) x �
;
4 4
� y; z � 4
.
�
1
9.8 kJ mol lower than the energy of two isolated molecules
of (I). This is an exaggerated estimation for the hydrogen-
bonding energy in this dimer, owing to the basis-set super-
position error.
The dimers in (I) form stacks along [001], with an inter-
Ê
planar distance of 3.446 (8) A between two neighbouring
Compound (II)
Crystal data
+
ÁCl�
C
M
9
H
7
N
2
O
3
Cu Kꢁ radiation
r
= 226.62
Cell parameters from 32
re¯ections
naphthalene fragments. It is likely that, under the effect of
stacking interactions, the dimers ¯atten in the crystal, at the
cost of distortion of the hydrogen-bonding geometry.
Monoclinic, P2 =c
1
Ê
ꢀ
a = 7.566 (4) A
ꢂ = 5.4±18.3
ꢃ = 3.61 mm
T = 295 (2) K
Ê
b = 7.467 (4) A
� 1
Ê
c = 16.334 (7) A
In (II), the cations form stacks along [100] (Fig. 2), with
Ê
interplanar distances of 3.341 (10) and 3.447 (10) A. The nitro
ꢀ
ꢅ = 92.49 (3)
V = 921.9 (8) A
Z = 4
Specimen shape: ¯at sheet
25 Â 25 Â 2.0 mm
Ê
3
ꢀ
Particle morphology: plate, yellow
group is twisted by 12.4 (5) with respect to the naphthalene
moiety. The anions make short contacts to NH, OH and CH
groups, which can be considered as hydrogen bonds (Table 2).
�
3
D
x
= 1.633 Mg m
Data collection
DRON-3M diffractometer (Bure-
vestnik, Russia)
Specimen mounting: pressed as a
thin layer in the specimen holder
Specimen mounted in re¯ection
mode
h = 0 ! 5
k = 0 ! 5
Experimental
l = � 12 ! 12
ꢀ
2ꢂmin = 5.0, 2ꢂmax = 70.0
Compound (I) was prepared by oxidation of 8-hydroxy-5-nitroso-
ꢀ
Increment in 2ꢂ = 0.02
quinoline by HNO
3
, followed by recrystallization from ethanol,
+
ÁCl�
ꢁ
o20 A. V. Yatsenko et al.
9 6 2
C H N O
3
9 7 2
and C H N O
3
Acta Cryst. (2002). C58, o19±o21

organic compounds
Table 2
Geometry of hydrogen bonds and short CÐHÁ Á ÁCl(O) contacts (A, ) for
II).
together into groups with a common Uiso parameter for each group.
The H atoms were placed in geometrically calculated positions and
their isotropic displacement parameters were ®xed. The planarity of
the naphthalene fragments and the nitro groups was restrained. The
anisotropy of the diffraction-line broadening was approximated by a
quartic form in hkl (Popa, 1998). The standard uncertainties obtained
from the Rietveld re®nement were corrected for serial correlation
effects (B e rar & Lelann, 1991).
Ê
ꢀ
(
DÐHÁ Á ÁA
DÐH
HÁ Á ÁA
DÁ Á ÁA
DÐHÁ Á ÁA
N1ÐH1Á Á ÁCl1
O8ÐH8Á Á ÁCl1
0.90
0.84
1.00
1.00
1.00
1.00
2.11
2.17
2.61
2.43
2.43
2.58
2.945 (7)
2.950 (6)
3.515 (8)
3.312 (10)
3.161 (10)
3.170 (10)
154
154
151
147
130
118
i
ii
C2ÐH2Á Á ÁCl1
iii
iv
v
C4ÐH4Á Á ÁO1
C6ÐH6Á Á ÁO2
C3ÐH3Á Á ÁO2
For both compounds, data collection: local program; cell re®ne-
ment: LSPAID (Visser, 1986); program(s) used to solve structure:
MRIA (Zlokazov & Chernyshev, 1992); program(s) used to re®ne
structure: MRIA; molecular graphics: PLATON (Spek, 2001); soft-
ware used to prepare material for publication: PARST (Nardelli,
1983).
1
2
3
2
Symmetry codes: (i) � x; 2 � y; 1 � z; (ii) � x; y �
;
� z; (iii) 1 � x; � y; 1 � z; (iv)
1
2
1
2
1
2
1
2
1
� x; ‡ y; � z; (v) x; � y; ‡ z.
Re®nement
R
R
R
p
= 0.040
wp = 0.056
exp = 0.030
Pro®le function: split-type pseudo-
Voigt
401 re¯ections
112 parameters
H-atom parameters constrained
(Á/ꢄ)max = 0.048
Preferred orientation correction:
spherical harmonics (Ahtee et al.,
1989)
Supplementary data for this paper are available from the IUCr electronic
archives (Reference: SK1513). Services for accessing these data are
described at the back of the journal.
S = 1.85
2
ꢀ
ꢂ
min = 9.40, 2ꢂmax = 70.0
ꢀ
Increment in 2ꢂ = 0.02
Wavelength of incident radiation:
1
Excluded region(s): 5.00±9.38
Ê
.5418 A
References
Ahtee, M., Nurmela, M., Suortti, P. & Jarvinen, M. (1989). J. Appl. Cryst. 22,
2
61±268.
The DFT and INDO±CISD calculations were performed using
programs provided by Dr D. N. Laikov (Laikov, 1997) and Professor
B. Dick (Dick & Nickel, 1983), respectively. The DFT calculations
were carried out using the BLYP (Becke±Lee±Yang±Parr) exchange-
correlation function (Becke, 1988; Lee et al., 1988) with a triple zeta
basis set, including polarization and diffuse functions for all atoms.
The orthorhombic and monoclinic cell dimensions of (I) and (II)
were determined with ITO (Visser, 1969) and TREOR90 (Werner et
al., 1985), respectively, and re®ned to M20 = 18 for (I) and 22 for (II),
and F = 43 for (I) and 39 for (II), using the ®rst 26 and 32 peak
positions, respectively. The initial molecular models were built with
MOPAC7 (Stewart, 1993) on the PM3 level (Stewart, 1989). In (I),
the position and orientation of the molecule were determined using
the grid-search procedure (Chernyshev & Schenk, 1998). In (II), the
position and orientation of the cation and the position of the anion
were determined using the simulated annealing procedure (Zhukov
et al., 2001). The X-ray diffraction pro®les and the difference
Becke, A. D. (1988). Phys. Rev. A, 38, 3098±3100.
B e rar, J.-F. & Lelann, P. (1991). J. Appl. Cryst. 24, 1±5.
Chernyshev, V. V. & Schenk, H. (1998). Z. Kristallogr. 213, 1±3.
Dick, B. & Nickel, B. (1983). Chem. Phys. 1, 1±16.
Ermakov, A. I., Voronin, V. G., Nelyubin, V. I., Oblapenko, E. I., Sorokin,
A. A., Epshtein, N. I. & Muravskaya, I. D. (1985). Khim. Farm. Zh. 19, 199±
204.
Laikov, D. N. (1997). Chem. Phys. Lett. 281, 151±156.
Lee, C., Yang, W. & Parr, R. G. (1988). Phys. Rev. B, 37, 785±789.
Nardelli, M. (1983). Comput. Chem. 7, 95±98.
Popa, N. C. (1998). J. Appl. Cryst. 31, 176±180.
Spek, A. L. (2001). PLATON. University of Utrecht, The Netherlands.
Stewart, J. J. P. (1989). J. Comput. Chem. 10, 209±220.
Stewart, J. J. P. (1993). MOPAC7. QCPE Program No. 455. Quantum
Chemistry Program Exchange, Department of Chemistry, Indiana Univer-
sity, Bloomington, Indiana, USA.
Visser, J. W. (1969). J. Appl. Cryst. 2, 89±95.
Visser, J. W. (1986). Powder Diffr. 1, 66±76.
Werner, P.-E., Eriksson, L. & Westdahl, M. (1985). J. Appl. Cryst. 18, 367±370.
Yatsenko, A. V., Paseshnichenko, K. A., Chernyshev, V. V. & Schenk, H.
(
2001). Acta Cryst. C57, 397±399.
Zhukov, S. G., Chernyshev, V. V., Babaev, E. V., Sonneveld, E. J. & Schenk, H.
2001). Z. Kristallogr. 216, 5±9.
B
re®nements are shown in Figs. 3 and 4. The ®nal R values were 0.067
for (I) and 0.076 for (II). The Cl atom was re®ned anisotropically; the
C, N and O atoms were re®ned isotropically and were gathered
(
Zlokazov, V. B. & Chernyshev, V. V. (1992). J. Appl. Cryst. 25, 447±451.
+ÁCl� o21
ꢁ
Acta Cryst. (2002). C58, o19±o21
A. V. Yatsenko et al.
C H N
9 6 2
O
3
and C H N
9 7 2
O
3