Synthesis, antimicrobial evaluation and QSAR study of some 3-hydroxypyridine-4-one and 3-hydroxypyran-4-one derivatives

Article abstract of DOI:10.1016/j.ejmech.2008.10.022

A series of Mannich bases of 2-alkyl-3-hydroxy-pyridine-4-ones, namely 2-alkyl-3-hydroxy-5-N-piperidylmethyl or N,N-dialkylaminomethyl pyridine-4-ones 9, 10 and 15-18, two derivatives of N-aryl-2-methyl-3-hydroxy-pyridine-4-ones 19, 20 and two N-alkyl derivatives of maltol, 21 and 22 were prepared. They were screened for their antibacterial and antifungal activities against a variety of microorganisms using micro plate Alamar Blue assay (MABA) method. Multiple linear regressions (MLR) analysis was performed for the synthesized compounds as well as a series of pyridinone and pyranone derivatives 23-43 which have been synthesized and evaluated for antimicrobial activity by other researchers previously. Studied compounds showed a better quantitative structure-activity relationship (QSAR) model for the antimicrobial activity against Candida albicans and Staphylococcus aureus in comparison with other tested microorganisms.

Full text of DOI:10.1016/j.ejmech.2008.10.022

European Journal of Medicinal Chemistry 44 (2009) 2145–2157
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis, antimicrobial evaluation and QSAR study of some
3-hydroxypyridine-4-one and 3-hydroxypyran-4-one derivatives
,
b
a
a
c
Afshin Fassihi a , Daryoush Abedi , Lotfollah Saghaie , Razieh Sabet , Hossein Fazeli ,
*
Ghasem Bostaki a, Omid Deilami a, Hekmatollah Sadinpour a
a Department of Medicinal Chemistry, Faculty of Pharmacy, Isfahan University of Medical Sciences, Hezar Jerib, 81746-73461 Isfahan, Islamic Republic of Iran
b Department of Biotechnology, Faculty of Pharmacy, Isfahan University of Medical Sciences, Hezar Jerib, 81746-73461 Isfahan, Islamic Republic of Iran
c Department of Microbiology, Faculty of Medicine, Isfahan University of Medical Sciences, Hezar Jerib, 81746-73461 Isfahan, Islamic Republic of Iran
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of Mannich bases of 2-alkyl-3-hydroxy-pyridine-4-ones, namely 2-alkyl-3-hydroxy-5-N-piper-
idylmethyl or N,N-dialkylaminomethyl pyridine-4-ones 9, 10 and 15–18, two derivatives of N-aryl-2-
methyl-3-hydroxy-pyridine-4-ones 19, 20 and two N-alkyl derivatives of maltol, 21 and 22 were
prepared. They were screened for their antibacterial and antifungal activities against a variety of
microorganisms using micro plate Alamar BlueÒ assay (MABA) method. Multiple linear regressions
(MLR) analysis was performed for the synthesized compounds as well as a series of pyridinone and
pyranone derivatives 23–43 which have been synthesized and evaluated for antimicrobial activity by
other researchers previously. Studied compounds showed a better quantitative structure–activity rela-
tionship (QSAR) model for the antimicrobial activity against Candida albicans and Staphylococcus aureus
in comparison with other tested microorganisms.
Received 9 July 2008
Received in revised form
15 October 2008
Accepted 20 October 2008
Available online 30 October 2008
Keywords:
3-Hydroxypyridin-4-ones
3-Hydroxypyran-4-ones
Antimicrobial activity
QSAR
Ó 2008 Elsevier Masson SAS. All rights reserved.
1. Introduction
achieved by sequestration of iron with iron-binding proteins, the
most abundant, haemoproteins which contain almost 80% of
It is almost 120 years that physicians have revealed that the
coincidence of blood and bacteria in a wound may cause a life-
threatening infection. It has also been shown that blood or
hemoglobin enhances the lethality of intraperitoneal or subcuta-
neous inocula of bacteria such as Escherichia coli. The effective
component of hemoglobin is iron, and various soluble iron
compounds exert an equivalent effect [1]. Increased susceptibility
to infectious disease is observed in iron-overloaded states such as
ß-thalassemia major and in hemolytic states such as sickle cell
disease. Administration of iron compounds to the host can increase
the virulence of E. coli, Listeria monocytogenes, Salmonella typhi-
murium and other pathogens [2]. In fact, iron is an essential
element required for the growth and virulence of virtually all
microbial pathogens [3,4]. Aerobic microorganisms need this
element for a variety of cellular functions including reduction of
oxygen for synthesis of ATP, replication of DNA, energy production
and protection of the cell against oxygen reactive species [5,6]. The
availability of iron is critically important in host-parasite interac-
tions [7,8]. Vertebrate hosts withhold iron from microbial invaders
as a major defence mechanism against infection [4,7,9]. This task is
the total iron of the vertebrates [10]. Some natural antibiotics,
called siderophores, have a sequestrating ability. They are low-
molecular-weight chelating agents that form stable complexes
with iron [11,12]. There are many reports of the antimicrobial
activity of chelating agents with different chemical structures
[13–19]. Kojic acid (5-hydroxy-2-hydroxymethyl-pyran-4-one) and
3-hydroxypyranones, derivatives of it are examples of these
compounds [13]. The bidentate chelating ligand 3-hydroxypyr-
anone, which has a catechol-like function, forms stable complexes
with several metal ions such as Fe3þ [20,21]. In vitro antibacterial
and antifungal activities of 3-hydroxypyridinones, bioisoster
derivatives of 3-hydroxypyranones with metal chelating ability
have been described. They have an inhibitory effect on the growth
of E. coli, Listeria inocua and Staphylococcus aureus [19,22,23].
Cephalosporins possessing 2-(5-hydroxy-4-pyridon-2-yl) ethenyl
moieties have shown strong activity against Pseudomonas aerugi-
nosa. Another cephalosporin derivative with a 1,5-dihydroxy-4-
pyridone-2-carboxamide moiety, MT0703, has excellent in vitro
and in vivo antibacterial activity against P. aeruginosa as well as
E. coli. The pyridinone ring in the structure of these cephalosporin
derivatives plays a significant role in anti-pseudomonas activity
[24,25]. More recently antibacterial and antifungal activities of
carboxamide derivatives of 3-hydroxypyranones, 5-hydroxypyr-
anones and 5-hydroxypyridinones have been reported [26,27].
* Corresponding author. Tel.: þ98311 7922562; fax: þ98 0311 6680011.
E-mail address: fassihi@pharm.mui.ac.ir (A. Fassihi).
0223-5234/$ – see front matter Ó 2008 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2008.10.022

2146
A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
Quantitative structure–activity relationship (QSAR) research
shown in Scheme 1. The commercially available maltol 1 or ethyl
maltol 2 were benzylated in 90% aqueous ethanol to give 3 and 4.
Refluxing 3 or 4 with aqueous ammonia in ethanol gave 2-alkyl-3-
benzyloxy pyridine-4-ones 5 and 6. The Mannich bases 7, 8 and 11–
14 were obtained by the reaction of 5 and 6 with 40% formaldehyde
solution and the appropriate secondary amine in ethanol. The
benzyl protection was removed by catalytic hydrogenation of 7, 8
and 11–14 to give the desired final Mannich bases 9, 10 and 15–18.
Synthesis of N-aryl-2-methyl-3-hydroxy-pyridine-4-ones 19 and 20
was achieved by refluxing maltol 1 with an excess of the suitable
primary aryl amines in acidic solution for 50–60 h (Scheme 2). The
N-alkyl derivatives of maltol, 21 and 22 (Scheme 3) were prepared
following the methodology as described by Harris from commer-
cially available maltol 1 [38]. Formation of the desired compounds
was confirmed on the basis of elemental analysis, IR, 1H NMR and
mass spectral data. Some of the characterization data of the
synthesized compounds are summarized in Table 1.
field has been widely developed because of its powerful ability to
predict drug activity [28]. It gives information that is useful for
molecular design and medicinal chemistry [29–33]. QSAR models
are mathematical equations relating chemical structures to their
biological activities as a linear regression model of the form
y ¼ Xb þ e. This equation may be used to describe a set of predictor
variables (X) with a predicted variable (y) by means of a regression
vector (b). In the first step of a typical QSAR study one needs to find
a set of molecular descriptors representing the higher impact on
the biological activity of interest. Multiple linear regression (MLR),
genetic algorithm (GA), partial least square (PLS) and principle
component analysis (PCA) are some of the variable selection
methods to build up such a set [34–37].
In the present paper, we describe the synthesis, structural
properties and antimicrobial activities of 6 novel Mannich bases of
2-alkyl-3-hydroxy-pyridine-4-ones as well as the antimicrobial
activity of 4 previously synthesized compounds that have this ring
in their structures. Their antimicrobial activities were examined by
micro plate Alamar BlueÒ assay method against various bacteria
and fungi. In the chemometrics part of this study we explore QSAR
model for a series of 3-hydroxypyridine-4-one and 3-hydrox-
ypyran-4-one derivatives which have been synthesized and eval-
uated for antimicrobial activity by Aytemir et al. as well as the
compounds prepared in this study [26,27].
2.2. Antimicrobial activity
Minimum inhibitory concentrations (MICs) were determined by
Micro plate Alamar BlueÒ Assay (MABA) method [39,40]. Tested
bacteria were three Gram-positive bacteria: S. aureus PTCC 1337,
Bacillus subtilis PTCC 1023 and L. monocytogenes PTCC 1165, three
Gram-negative bacteria: E. coli PTCC 1338, P. aeruginosa PTCC 1074
and Salmonella enteritidis PTCC 1091 obtained from Persian type
culture collection. The Muller–Hinton broth was used for bacteria
[41]. All compounds were screened for their antifungal activity
against one yeast-like fungus: Candida albicans PTCC 5027 and two
molds: Aspergillus niger 5021 and Aspergillus flavus PTCC 5003
obtained from Persian type culture collection. RPMI 1640 medium
buffered with MOPS at pH 7.0 was used for fungi [41]. The anti-
bacterial and antifungal activity data are given in Table 2.
Our QSAR analysis establishes mathematical relationship
between biological activities and computable parameters such as
topological, physicochemical, stereochemical or electronic indices.
2. Results and discussion
2.1. Chemistry
Mannich bases of 2-alkyl-3-hydroxy-pyridine-4-ones, namely 2-
alkyl-3-hydroxy-5-N-piperidylmethyl or N,N-dialkylaminomethyl
pyridine-4-ones 9,10 and 15–18, wereprepared by the methodology
The investigation of antimicrobial screening data revealed that
some of the tested compounds showed moderate to good bacterial
growth inhibition. Compounds 9, 20 and 21 inhibited S. aureus
O
O
O
O
R
OBz
OBz
OBz
OH
N
NH4OH
CH2O
RR'NH
BzCl
'
R
N
H
R2
O
R2
N
H
R2
O
R2
11. R2 = CH3 , R = R'= CH3
1. R2 = CH3
2. R2 = C2H5
3. R2 = CH3
4. R2 = C2H5
5. R2 = CH3
6. R2 = C2H5
'
12. R2 = C2H5 , R = R= CH3
13. R2 = CH3 , R = R'= C2H5
'
Pd/H2
Piperidine
14. R2 = C2H5 , R = R= C2H5
CH2O
O
O
R
OH
R2
OBz
R2
N
N
'
R
N
H
N
H
15. R2 = CH3 , R = R'= CH3
7. R2 = CH3
8. R2 = C2H5
16. R2 = C2H5 , R = R'= CH3
17. R2 = CH3 , R = R'= C2H5
18. R2 = C2H5 , R = R'= C2H5
Pd/H2
O
OH
N
9. R2 = CH3
10. R2 = C2H5
N
H
R2
Scheme 1. Synthetic methodology adopted for the preparation of Mannich bases (9 and 10) and (15–18).

A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
2147
O
could interact with receptor more efficiently. This could propose
another mechanism; a receptor mediated one, if R2 and Q2 had
significant values for the antimicrobial activity of these compounds.
The second equation of Table 8 was found by using constitu-
tional descriptors (E2). This equation explained the positive effect of
number of double bonds (nDB) and number of rotatable bonds
(RBN) on the antimicrobial activity of studied compounds.
The equation E3 of Table 8 obtained from the pool of topological
descriptors, explained the negative effect of average connectivity
index chi-1 (X1A) and molecular multiple path count of order 7
(piPC07) on antimicrobial activity of the compounds.
O
NH2
OH
OH
Reflux in EtOH
pH= 5
+
O
CH3
N
CH3
X
1
19: X: H,
20: X:m-OH
X
The equation E4 of Table 8 was found by using geometrical
descriptors. This equation explained the positive effect of molecular
electrotopological variation (DELS) and the negative effect of 3D
petijean shape index (PJI3) on the antimicrobial activity of
compounds.
Scheme 2. Synthesis of N-aryl-3-hydroxypyridine-4-ones (19 and 20) via the single
step synthetic pathway.
The effect of functional groups on the antimicrobial activity of
the studied compounds has been described by equation E5 of
Table 8. This equation explained the positive effect of the number
of secondary amides (aliphatic) (nCONHR) on the antimicrobial
activity of the studied compounds. The positive sign of the coef-
ficient of nCONHR proposed that an increase in the number of
secondary amides (aliphatic) resulted in enhanced activity.
The last equation (E6) was obtained from the all calculated
descriptors. Stepwise selection and elimination of variables
produced a three-parametric QSAR equation. In this model DMy
and nDB have a positive effect and PJI3 has a negative effect on
inhibitory activity.
growth at 16 mg/ml concentration. Compound 9 inhibited S. enter-
itidis at the same concentration. It was also a moderate inhibitor of
B. subtilis.
2.3. QSAR analysis
The biological data used in this study are antimicrobial activity,
(in terms of ꢀlog MIC), of a set of ten 3-hydroxypyridine-4-one
derivatives 9, 10, 15–22 which we prepared and twenty-one
derivatives of 3-hydroxypyridine-4-one and 3-hydroxypyran-4-
one 23–43 which have been reported before as antimicrobial
agents [26,27].
In Table 9 the resulted equations for all compounds against
C. albicans are listed. The first equation of Table 9 was found by
using quantum descriptors (E1). It explained the positive effect of
SUMPC and softness on antimicrobial activity of the compounds.
The positive coefficient of SUMPC reveals the presence of columbic
interactions between the ligands and receptors and demonstrates
that ligands with most SUMPC could interact with receptor more
efficiently. This could propose again the receptor mediated mech-
anism for the antimicrobial activity of these compounds if R2 and Q2
had significant values. The second equation of Table 9 was found by
using chemical descriptors (E2). It shows the positive effect of mass
and the negative effect of surface area (SA) on the antimicrobial
activity of the compounds. The negative coefficient of surface area
indicates that increasing this parameter hinders the ligand to pass
through the cell membrane and thus decreases the activity.
The effect of constitutional descriptors on the antimicrobial
activity of the studied compounds has been described by equation
E3 of Table 9. It explained the positive effect of mean atomic
Sanderson electronegativity (scaled on Carbon atom) (Me) and
number of multiple bonds (nBM) on the antimicrobial activity. The
MLR equation of Table 9 was obtained from the pool of topological
descriptors (E4). It includes the positive effect of Eigenvalue sum
from Z weighted distance matrix (Barysz matrix) (SEigz) and of
distance/detour ring index of order 10 (D/Dr10) and the negative
effect of Balaban centric index (BAC) on the antimicrobial activity.
The equation obtained from the effect of geometrical parameters on
the antimicrobial activity of the studied compounds (E5), shows
The structural features of these compounds are listed in Table 3.
The calculated descriptors for each molecule are summarized in
Table 4. The antimicrobial activities are summarized in Tables 5–7
and then were used for subsequent QSAR analysis as dependent
variables.
Separate stepwise selection-based MLR analyses were per-
formed using different types of descriptors, and then, an MLR
equation was obtained utilizing the pool of all calculated descrip-
tors. The results are summarized in Tables 8–14. Correlation coef-
ficient (r2) matrices for the descriptors used in different MLR
equations are shown in Tables 15–21. Collinear descriptors degrade
the performance of MLR equations and such models have lowered
prediction ability. Tables 15–21 show that no significant correlation
exists between pairs of descriptors.
In Tables 8–14 the MLR analysis with different types of
descriptors of tested compounds against S. aureus, C. albicans and
P. aeruginosa are listed.
Table 8 provides the resulted equations for all of the compounds
against S. aureus. In this series the chemical parameter did not have
a significant impact on the antimicrobial activity. The equation E1
shows that among quantum descriptors, MPC and DMy have
a positive effect on the antimicrobial activity; this contribution
suggests that electronic interaction plays an important role in
inhibitory activity of these compounds. The positive coefficient of
MPC reveals the presence of columbic interactions between the
ligands and receptors. According to equation E1 a negative region in
receptor produces columbic interaction; ligands with most MPC
O
O
O
O
OBn
CH3
OH
OBn
CH3
OH
H2
RNH2
NaOH/H2O
Bz-Cl
Pd/C
N
R
O
CH3
O
N
R
CH3
21: R: C3H7
22: R: C4H9
1
Scheme 3. Synthesis of N-alkyl-2-methyl-3-hydroxypyridine-4-ones 21 and 22 utilizing the methodology of Harris.

2148
A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
Table 1
Characterization data of the synthesized compounds.
O
X
R5
OH
R2
No.
X
R2
R5
Mol. formula
M.p (ꢁC)
Yield (%)
Analysis (%)
Found (Calculated)
C
H
N
9
NH
NH
NH
NH
NH
NH
N–Ph
N-m-OH–Ph
CH3
C2H5
CH3
C2H5
CH3
C2H5
CH3
CH3
CH2–Ra
C12H18N2O2
C13H20N2O2
C9H14N2O2
C10H16N2O2
C11H18N2O2
C12H20N2O2
C12H11NO2
C12H11NO3
269–270
270–272
252–253
254–255
252–253
256–257
221–222
268–269
22
25
25
30
23
23
35
40
64.67 (64.84)
65.96 (66.07)
59.11 (59.32)
60.06 (61.20)
62.81 (62.83)
64.11 (64.26)
71.55 (71.63)
66.12 (66.35)
8.12 (8.16)
8.32 (8.53)
7.70 (7.74)
8.25 (8.22)
8.59 (8.63)
8.86 (8.99)
5.35 (5.51)
4.97 (5.10)
12.62 (12.60)
11.83 (11.85)
15.43 (15.37)
14.35 (14.27)
13.43 (13.32)
12.46 (12.49)
7.08 (6.96)
10
15
16
17
18
19
20
CH2–Ra
CH2–N(CH3)2
CH2–N(CH3)2
CH2–N(C2H5)2
CH2–N(C2H5)2
H
H
6.54 (6.45)
Ra is
.
N
a negative effect of 3D Balaban centric index (J3D) and asphericity
(ASP) on the antimicrobial activity.
(nR10). The positive sign of the coefficient of the nR10 proposes
that increasing the number of nR10 of the molecule results in an
enhanced activity. The MLR equation of Table 10 obtained from the
pool of topological descriptors (E3) explained the positive effect of
Eigenvalue sum from Z weighted distance matrix and the negative
effect of information content index (neighborhood symmetry of 3-
order) (IC3). The effect of geometrical descriptors on the antimi-
crobial activity of the studied compounds has been described by
equation E4 of Table 10. It explains the positive effect of maximal
electrotopological negative variation (MAXDN) on the activity of
these compounds. Equation E5 shows that among the functional
group descriptors, nNR2 has a negative effect on the antimicrobial
activity of these compounds. It shows that decreasing the number
of nNR2 results in enhanced activity. The equation E6 was obtained
from the pool of all calculated descriptors and it was similar to
equation E3.
Table 11 lists the resulted equation for the antimicrobial activity
of compounds 9, 10, 15–24 against C. albicans. The MLR equation of
Table 11 was obtained from the quantum descriptors (E1). It shows
the negative effect of highest-occupied molecular orbital (HOMO)
on the antimicrobial activity of compounds. This contribution
suggests that electronic interaction plays important roles in the
antimicrobial activity of these compounds. Molecules with low
HOMO energy values are more able to accept electrons than
The MLR equation of Table 9 obtained from the pool of func-
tional group descriptors (E6) explained the positive effect of the
number of secondary amides (aliphatic) (nCONHR) and the number
of substituted aromatic C (SP2) (nCaR) and the negative effect of the
number of tertiary amines (aliphatic) (nNR2) on the antimicrobial
activity. It shows that a decrease in the number of nNR2 and an
increase in the number of nCONHR and nCaR result in enhanced
activity. The last equation E7 was derived from the pool of all
calculated descriptors. It shows the negative effect of J3D, ASP and
superpendentic index (SPI) and the positive effect of the number of
total secondary C (SP3) (nCs) on the antimicrobial activity. This
equation, which has a high statistical quality could explain and
predict 0.81% and 0.73% of variance in pMIC data, respectively.
Table 10 describes that the QSAR models for the antimicrobial
activity against P. aeruginosa resulted for all compounds by using
different sets of molecular descriptors. The first equation (E1) was
found by using quantum descriptors and includes the negative
effect of hardness on antimicrobial activity of compounds. The
effect of constitutional descriptors on the antimicrobial activity of
the studied compounds has been described by equation E2 of Table
10. It shows the positive effect of average molecular weight (AMW),
rotatable bond fraction (RBF) and number of 10-membered rings
Table 2
Antibacterial and antifungal activities of synthesized compounds.
Compound
MIC mg/ml
E. coli
S. enteritidis
P. aeruginosa
S. aureus
L. monocytogenes
B. subtilis
C. albicans
A. niger
A. flavus
a
9
10
15
16
17
18
19
20
21
22
23
24
64
512
512
512
512
512
128
512
128
512
–
16
512
512
256
128
512
128
64
512
256
256
256
256
256
512
512
512
512
–
16
–
–
64
–
512
512
–
512
512
–
512
–
32
512
–
–
–
512
256
–
256
512
–
128
128
–
512
512
512
–
512
256
256
512
128
2c
512
512
512
512
512
512
–
512
512
256
128
512
4c
–
256
–
512
512
-
128
16
16
256
–
512
512
512
–
256
256
256
16
128
256
–
–
512
1b
–
–
256
0.5b
–
–
512
8c
Standard
0.25b
1b
0.5b
0.06b
a
Indicates bacteria are resistant to the compounds >512
mg/ml; MIC (
mg/ml) ¼ minimum inhibitory concentration, i.e., the lowest concentration to completely inhibit
bacterial growth.
b
Ciprofloxacin.
Ketoconazole.
c

A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
2149
Table 3
Chemical structure of the compounds used in QSAR analysis.
O
X
R5
R6
R3
R2
Compound
X
R2
R3
R5
R6
9
NH
NH
CH3
C2H5
OH
OH
CH2–Ra
H
H
10
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
CH2–Ra
NH
NH
NH
NH
N-Ph
N-m-OH-Ph
N-C3H7
N-C4H9
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
CH3
C2H5
CH3
C2H5
CH3
CH3
CH3
CH3
CH2Cl
CH3
CH2OH
CH2OH
CHO
COOH
CONHRb
CONHRc
CONHRd
CONHRb
CONHRc
CONHRd
CH2OH
COOH
CONHPh
CONHPh
CONHPh
CONH–Re
CONH–Re
CONH–Re
CH2OH
OH
OH
OH
OH
OH
OH
OH
OH
H
H
OH
OCH2Ph
OCH2Ph
OCH2Ph
OCH2Ph
OCH2Ph
OCH2Ph
OH
OH
OH
H
H
H
H
H
H
CH2–N(CH3)2
CH2–N(CH3)2
CH2–N(C2H5)2
CH2–N(C2H5)2
H
H
H
H
OH
OH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
OCH2Ph
OCH2Ph
OCH2Ph
OCH2Ph
OH
OCH2Ph
OCH2Ph
OH
N–CH3
N–CH3
O
N–CH3
N–CH3
O
H
H
H
OH
CH3
N
.
b is
Ra is
N
,
c
, Rd is
,
R
e is
,
R
R is
O
O
N
molecules with high HOMO energy values. The effect of chemical
descriptors on the antimicrobial activity of the studied compounds
has been described by equation E2 of Table 11. It explained the
negative effect of surface area (SA) on the antimicrobial activity of
the compounds. The negative coefficient of surface area indicates
that an increase in the SA of the molecule hinders the ligand to pass
through cell membrane and thus decreases the activity. Equation
E3, which was derived from a pool of constitutional constants,
shows the positive effect of number of oxygen atoms (nO) on the
antimicrobial activity. The positive sign of the coefficient of the nO
Table 4
Brief description of some descriptors used in this study.
Descriptor type
Constitutional
Molecular description
Molecular weight, no. of atoms, no. of non-H atoms, no. of bonds, no. of heteroatoms, no. of multiple bonds (nBM), no. of aromatic bonds, no.
of functional groups (hydroxy, amine, aldehyde, carbonyl,nitro, nitroso, ..), no. of rings, no. of circuits, no of H-bond donors, no of H-bond
acceptors, no. of Nitrogen atoms (nN), chemical composition, some of Kier–Hall electrotopological states (Ss), mean atomic polarizability (Mp),
number of rotable bonds (RBN), mean atomic sanderson electronegativity (Me),
Topological indices
Molecular size index, molecular connectivity indices (X1A, X2v, X2Av, X3Av, X4Av), information content index (IC), Kier Shape indices, total
walk count, path/walk-Randic shape indices, Zagreb indices, Schultz indices, Balaban J index (such as MSD) Wiener indices, topological charge
indices, Sum of topological distances between F.F (T(F.F), Ratio of multiple path count to path count (PCR), Mean information content
vertex degree magnitude (IVDM), Eigenvalue sum of Z weighted distance matrix (SEigz), reciprocal hyper-detour index (Rww), Eigenvalue
coefficient sum from adjacency matrix (VEA1)
Geometrical
Quantum
3D petijean shape index (PJI3), Gravitational index, Balaban index, Wiener index,
Highest-occupied Molecular Orbital Energy (HOMO), Lowest Unoccupied Molecular Orbital Energy (LUMO), Most positive charge (MPC), Least
negative charge (LNC), Sum of squares of charges (SSC), Sum of square of positive charges (SSPC), Sum of square of negative charges (SSNC), Sum
of positive charges (SUMPC), Sum of negative charges (SUMNC), Sum of absolute of charges (SAC), Total dipole moment (DMt), Molecular dipole
moment at X-direction (DMx), Molecular dipole moment at Y-direction (DMY), Molecular dipole moment at Z-direction (DMZ), Electronegativity
(
c
¼ ꢀ0.5(HOMO ꢀ LUMO), Electrophilicity (
u
¼
c
2/2
h), Hardness (
h
¼ 0.5(HOMO þ LUMO)), Softness (S ¼ 1/
h)
Functional group
Number of total tertiary carbons (nCt), Number of H-bond acceptor atoms (nHAcc), number of total hydroxyl groups (nOH), number of
unsubstituted aromatic C(nCaH), number of ethers (aromatic) (nRORPh)
Chemical descriptors
log P, Hydration Energy (HE), Polarizability (Pol), Molar refractivity (MR), Molecular volume (V), Molecular surface area (SA)

2150
A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
Table 5
Table 7
Experimental and predicted activity of compounds against Staphylococcus aureus.
Experimental and predicted activity of compounds against Pseudomonas aeruginosa.
Compound
Experimental pMIC
Predicted pMIC
REP (%)
0.917309
Compound
Experimental pMIC
Predicted pMIC
REP(%)
9
3.29
3.29
3.29
3.29
4.19
3.29
3.89
3.29
3.29
3.89
3.59
3.59
3.59
3.59
4.19
3.59
3.59
3.59
3.89
4.19
3.59
5.1
3.59
3.59
3.89
3.89
4.8
3.89
3.59
4.49
3.59
3.320549
3.300702
3.226554
3.397621
3.749766
3.320459
3.825538
3.26982
9
3.59
3.59
3.59
3.59
3.29
3.59
3.29
3.29
3.89
3.29
4.19
3.59
3.89
3.59
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.59
3.59
3.89
3.59
3.89
3.59
3.89
3.443974
3.530871
3.472232
3.636632
3.510631
3.587978
3.723022
3.747837
3.653376
3.599309
3.795735
3.785079
3.76962
3.715849
3.742194
3.782049
3.888874
3.900399
3.897727
3.837207
3.873711
3.921648
3.744965
3.815239
3.682614
3.578922
3.708554
3.68836
ꢀ4.24004
ꢀ1.67462
ꢀ3.3917
10
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
0.324245
ꢀ1.96638
3.167543
ꢀ11.7403
0.917309
ꢀ1.68504
ꢀ0.61715
ꢀ0.04402
0.973778
0.846993
3.634035
ꢀ2.38832
0.862738
ꢀ0.81191
ꢀ0.81226
0.764741
ꢀ0.99153
0.129311
ꢀ2.20787
3.278718
ꢀ0.31399
ꢀ1.03178
3.553385
0.821364
2.209202
0.045312
ꢀ0.80105
ꢀ2.84698
0.454943
ꢀ3.37462
10
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
1.282281
6.284661
ꢀ0.05636
11.63093
12.21603
ꢀ6.47687
8.593572
ꢀ10.387
5.153899
ꢀ3.19344
3.386829
ꢀ3.94973
ꢀ2.8543
ꢀ0.02894
0.266625
0.198251
ꢀ1.37581
ꢀ0.42049
0.807
3.288552
3.928252
3.620667
3.725382
3.506259
3.621242
4.156255
3.561075
3.617666
3.554753
3.895037
4.099489
3.711696
5.084036
3.553337
3.722266
3.922216
3.977879
4.802176
3.859087
3.490623
4.51052
ꢀ3.8728
ꢀ1.95954
2.514894
ꢀ0.30954
ꢀ4.89264
2.666772
ꢀ7.52046
3.955999
ꢀ2.37908
3.617916
3.73787
3.799604
3.472806
Table 6
Experimental and predicted activity of compounds against Candida albicans.
proposes that an increase in the number of (nO) of the molecule
resulted in activity enhancement. The equation E4 of Table 11 was
found by using topological descriptors. This equation explained the
positive effect of Balban centric index (BAC) on antimicrobial
activity of compounds 9, 10, 15–24. The effect of geometrical
descriptors on the antimicrobial activity of these compounds has
been described by equation E5 of Table 11. It explained the negative
effect of 3D petijean shape index (PJI3) and positive effect of sum of
geometrical distances between N.N, i.e., G(N.N) on the antimi-
crobial activity. The effect of functional groups on the antimicrobial
activity of these compounds has been described by equation E6 of
Table 11. This equation describes the structure–activity relationship
better than those obtained from the quantum, chemical, constitu-
tional and topological descriptors. The two-parametric QSAR
equation has correlation coefficient and standard error equal to
0.81 and 0.13, respectively. It explained the negative effect of
number of total primary C (SP3) (nCp) and number of unsubstituted
aromatic C (SP2) (nCaH) on the antimicrobial activity of the
compounds. The last equation E7 was obtained from the all calcu-
lated descriptors. Stepwise selection and elimination of variables
produced again a two-parametric functional group QSAR equation.
This equation could explain and predict 0.81% and 0.61% of variance
in pMIC data, respectively.
Compound
Experimental pMIC
Predicted pMIC
REP(%)
3.630409
10
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
35
36
37
38
39
40
41
42
43
3.29
3.29
3.89
3.29
3.29
3.59
3.29
3.59
3.89
3.89
3.89
3.89
3.89
3.89
4.49
4.49
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.89
3.41394
3.46932
3.892
5.168736
0.051387
2.057682
2.763123
ꢀ1.19289
0.927189
ꢀ7.82108
6.773426
ꢀ3.78703
0.492167
ꢀ4.91906
ꢀ5.44263
ꢀ1.24329
ꢀ0.53514
0.391336
ꢀ4.91906
ꢀ2.32964
1.581278
ꢀ3.87266
2.642907
2.675547
ꢀ1.06915
ꢀ3.53041
1.526217
2.661909
0.200625
3.525893
3.35912
3.38349
3.54768
3.32079
3.32959
4.17263
3.74806
3.90924
3.70762
3.68921
3.84223
4.4661
4.50764
3.70762
3.80144
3.9525
3.74497
3.9956
3.99694
3.84885
3.75735
3.95029
3.99638
3.89782
4.03217
In Table 12 the resulted equation for the antimicrobial activity
of compounds 9, 10, 15–24 against P. aeruginosa are listed. It should
Table 8
The results of MLR analysis with different types of descriptors of all compounds (Staphylococcus aureus).
No.
Descriptor source
MLR equations
N
R2
S.E
RMSCV
Q2
F
E1
E2
E3
E4
E5
E6
Quantum
Constitutional
Topological
Geometrical
Functional group
Molecular descriptor
pIC50 ¼ 2.505(ꢂ0.647) þ 3.507(ꢂ1.691) MPC þ 0.121(ꢂ0.048)DMy
pIC50 ¼ 3.775(ꢂ0.254) þ 0.148(ꢂ0.036)nDB ꢀ 0.115(ꢂ0.049)RBN
pIC50 ¼ 20.128(ꢂ4.595) ꢀ 35.306(ꢂ9.853)X1A ꢀ 0.001(ꢂ0.001)piPCO7
pIC50 ¼ 4.531(ꢂ0.570) þ 0.031(ꢂ0.009)DELS ꢀ 1.859(ꢂ0.688) PJI3
pIC50 ¼ 3.605(ꢂ0.081) þ 0.521(ꢂ0.151)nCONHR
31
31
31
31
31
31
0.45
0.41
0.34
0.37
0.30
0.55
0.34
0.36
0.37
0.36
0.38
0.50
0.37
0.42
0.39
0.40
0.40
0.36
0.31
0.18
0.23
0.21
0.17
0.35
11.52
8.74
7.20
8.51
11.95
11.24
pIC50 ¼ 4.544(ꢂ0.495) þ 0.130(ꢂ0.040)DMy þ 0.082(ꢂ0.031)nDB ꢀ 1.275(ꢂ0.584)PJI3

A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
2151
Table 9
The results of MLR analysis with different types of descriptors of all compounds (Candida albicans).
No. Descriptor source
MLR equations
N
R2
S.E
RMSCV Q2
F
E1
E2
E3
E4
E5
E6
E7
Quantum
Chemical
Constitutional
Topological
Geometrical
Functional group
pIC50 ¼ 5.727(ꢂ0.572) þ 0.391(ꢂ0.093)Softness þ 0.117(ꢂ0.049)SUMPC
pIC50 ¼ 4.241(ꢂ0.252) þ 0.006(ꢂ0.001)Mass ꢀ 0.005(ꢂ0.001)SA
28 0.50 0.22 0.24
28 0.62 0.19 0.21
28 0.62 0.19 0.21
28 0.73 0.16 0.19
28 0.51 0.18 0.19
28 0.67 0.18 0.20
0.35 12.05
0.52 20.65
0.51 20.96
0.60 22.16
0.60 25.88
0.53 16.16
0.73 25.30
pIC50 ¼ ꢀ3.598(ꢂ2.072) þ 0.0.039(ꢂ0.007)nBM þ 6.897(ꢂ2.019)ME
pIC50 ¼ 3.724(ꢂ0.190) þ 0.003(ꢂ0.001)D/Dr10 ꢀ 0.023(ꢂ0.005)BAC þ 0.360(ꢂ0.160)SEigz
pIC50 ¼ 5.173(ꢂ0.202) ꢀ 0.370(ꢂ0.056)J3D ꢀ 0.983(ꢂ0.301)ASP
pIC50 ¼ 3.781(ꢂ0.065) ꢀ 0.312(ꢂ0.077)nNR2 þ 0.247(ꢂ0.078)nCONHR þ 0.095(ꢂ0.032)nCaR
Molecular descriptor pIC50 ¼ 6.192(ꢂ0.292) ꢀ 0.714(ꢂ0.097)J3D ꢀ 1.519(ꢂ0.269)ASP þ 0.249(ꢂ0.065)nCs ꢀ 0.001(ꢂ0.00)SPI 28 0.82 0.14 0.15
Table 10
The results of MLR analysis with different types of descriptors of all compounds (Pseudomonas aeruginosa).
No.
Descriptor source
MLR equations
N
R2
S.E
RMSCV
Q2
F
E1
E2
E3
E4
E5
E6
Quantum
Constitutional
Topological
Geometrical
Functional group
Molecular descriptor
pIC50 ¼ 2.177(ꢂ0.430) ꢀ 8.999(ꢂ2.507)Hardness
31
31
31
31
31
31
0.55
0.62
0.53
0.21
0.35
0.53
0.19
0.14
0.16
0.20
0.18
0.16
0.19
0.19
0.17
0.21
0.19
0.17
0.25
0.25
0.44
0.13
0.26
0.44
12.88
14.76
16.11
7.94
15.69
16.11
pIC50 ¼ 1.870(ꢂ0.314) þ 0.172(ꢂ0.030)AMW þ 2.720(ꢂ0.753)RBF þ 0.175(ꢂ0.071)nR10
pIC50 ¼ 4.138(ꢂ0.347) þ 0.807(ꢂ0.143)SEigz ꢀ 0.288(ꢂ0.096)IC3
pIC50 ¼ 2.859(ꢂ0.307) þ 0.396(ꢂ0.141)MAXDN
pIC50 ¼ 3.839(ꢂ0.046) ꢀ 0.270(ꢂ0.068)nnr2
pIC50 ¼ 4.138(ꢂ0.347) þ 0.807(ꢂ0.143)SEigz ꢀ 0.288(ꢂ0.096)IC3
Table 11
The results of MLR analysis with different types of descriptors of compounds 9, 10, 15–24 (Candida albicans).
No.
Descriptor source
MLR Equations
N
R2
S.E
RMSCV
Q2
F
E1
E2
E3
E4
E5
E6
E7
Quantum
Chemical
Constitutional
Topological
Geometrical
pIC50 ¼ ꢀ0.16(ꢂ1.353) ꢀ 11.201(ꢂ4.207)HOMO
10
10
10
10
10
10
10
0.46
0.46
0.42
0.49
0.78
0.81
0.81
0.21
0.22
0.23
0.25
0.24
0.20
0.18
0.18
0.32
0.31
0.19
0.27
0.43
0.61
0.61
6.89
6.87
5.87
pIC50 ¼ 4.891(ꢂ0.523) ꢀ 0.004(ꢂ0.001)SA
0.21
0.22
0.20
0.14
0.13
0.13
pIC50 ¼ 2.674(ꢂ0.360) þ 0.373(ꢂ0.154)nO
pIC50 ¼ 3.943(ꢂ0.162) ꢀ 0.022 (ꢂ0.008)BAC
7.76
pIC50 ¼ 6.213(ꢂ0.544) ꢀ 3.633(ꢂ0.748)PJI3 þ 0.095(ꢂ0.034)G(N.N)
pIC50 ¼ 4.014(ꢂ0.098) ꢀ 0.244(ꢂ0.046)nCp ꢀ 0.076(ꢂ0.025)nCaH
pIC50 ¼ 4.014(ꢂ0.098) ꢀ 0.244(ꢂ0.046)nCp ꢀ 0.076(ꢂ0.025)nCaH
12.90
14.98
14.98
Functional group
Molecular descriptor
Table 12
The results of MLR analysis with different types of descriptors of compounds 9, 10, 15–24 (Pseudomonas aeruginosa).
No.
Descriptor source
MLR Equations
N
R2
S.E
RMSCV
Q2
F
E1
E2
E3
Quantum
Topological
Molecular descriptor
pIC50 ¼ 0.167(ꢂ1.381) ꢀ 10.775(ꢂ4.373)HOMO
12
12
12
0.37
0.72
0.72
0.22
0.15
0.15
0.18
0.18
0.18
0.28
0.54
0.54
6.07
11.50
11.50
pIC50 ¼ 5.70(ꢂ0.452) ꢀ 0.855(ꢂ0.182)IDDE þ 0.190(ꢂ0.072)PHI
pIC50 ¼ 5.70(ꢂ0.452) ꢀ 0.855(ꢂ0.182)IDDE þ 0.190(ꢂ0.072)PHI
Table 13
The results of MLR analysis with different types of descriptors of compounds 25–43 (Staphylococcus aureus).
No.
Descriptor source
MLR equations
N
R2
S.E
RMSCV
Q2
F
E1
E2
E3
E4
Quantum
Constitutional
Topological
pIC50 ¼ 3.927(ꢂ0.081) þ 0.731(ꢂ0.049)DMy
pIC50 ¼ 3.477(ꢂ0.203) þ 0.095(ꢂ0.040) nDB
pIC50 ¼ 3.412(ꢂ0.183) þ 0.035 (ꢂ0.011)BAC
pIC50 ¼ 4.509(ꢂ0.143) ꢀ 0.422 (ꢂ0.152)
19
19
19
19
0.42
0.25
0.35
0.61
0.35
0.40
0.37
0.30
0.39
0.19
0.19
0.36
0.24
0.15
0.26
0.40
12.63
5.68
9.45
7.89
Functional group
nRORPh ꢀ 0.626(ꢂ0.175)nCs ꢀ 0.159(ꢂ0.062)nCaR
pIC50 ¼ 8.241(ꢂ0.820) þ 0.124(ꢂ0.023)DMy0.103
(ꢂ0.120)nAB ꢀ 21.398(ꢂ3.590)RBF ꢀ 0.177(ꢂ0.034)L/BW ꢀ 0.130
(ꢂ0.035)VRA2 þ 0.204(ꢂ0.089)SPAN
E5
Molecular descriptor
19
0.93
0.15
0.19
0.82
25.48
Table 14
The results of MLR analysis with different types of descriptors of compounds 25–33, 35–43 (Candida albicans).
No.
Descriptor source
MLR equations
N
R2
S.E
RMSCV
Q2
F
E1
E2
E3
Chemical
Constitutional
Topological
pIC50 ¼ 3.602(ꢂ0.167) þ 0.001(ꢂ0.001)Mass
pIC50 ¼ 3.769(ꢂ0.055) þ 0.027(ꢂ0.006)nAB
pIC50 ¼ ꢀ5.047(ꢂ2.249) þ 0.002(ꢂ0.000)
piPC07 þ 30.502(ꢂ7.436)X2A þ 0.124(ꢂ0.043)SEigz
pIC50 ¼ 3.167(ꢂ0.324) þ 0.160(ꢂ0.065)MAXDP
pIC50 ¼ 3.817(ꢂ0.044) þ 0.111(ꢂ0.024)nCaR
pIC50 ¼ 4.026(ꢂ0.072) þ 0.167(ꢂ0.024)
18
18
18
0.23
0.53
0.81
0.17
0.13
0.09
0.20
0.16
0.10
0.14
0.29
0.71
7.60
18.21
20.47
E4
E5
E6
Geometrical
Functional group
Molecular descriptor
18
18
18
0.27
0.56
0.82
0.17
0.13
0.08
0.18
0.16
0.12
0.16
0.30
0.61
20.62
20.62
21.94
nCaR ꢀ 0.588(ꢂ0.164) ASP ꢀ 0.237(ꢂ0.081)nNHRPh

2152
A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
Table 15
Correlation coefficient (r2) matrix for the descriptors of all compounds used in the different MLR equations (Staphylococcus aureus).
MPC
1
DMy
nDB
RBN
X1A
piPCO7
DELS
0.752
PJI3
0.190
nCONHR
pMIC
0.572
MPC
DMy
nDB
RBN
X1A
piPCO7
DELS
PJI3
nCONHR
pMIC
0.540
1
0.545
0.368
1
0.016
0.022
0.377
1
ꢀ0.764
ꢀ0.568
ꢀ0.512
ꢀ0.053
1
0.633
0.418
0.331
0.243
ꢀ0.631
1
0.675
0.404
0.652
0.288
ꢀ0.528
0.369
0.608
0.016
1
0.488
0.677
0.317
ꢀ0.616
0.780
1
ꢀ0.070
0.016
0.522
ꢀ0.191
0.310
0.190
1
0.606
0.531
ꢀ0.074
ꢀ0.466
0.192
0.465
ꢀ0.307
0.540
1
Table 16
Correlation coefficient (r2) matrix for the descriptors of all compounds used in the different MLR equations (Candida albicans).
SUMPC Softness Mass SA
nBM
Me
BAC
0.030
0.761 ꢀ0.597
SEigz
0.504
0.702
0.663
0.365
0.574
D/Dr10 J3D
0.664 ꢀ0.468 ꢀ0.060
0.180
0.788 ꢀ0.001 ꢀ0.791
0.725 ꢀ0.633 ꢀ0.075 ꢀ0.213
0.517 ꢀ0.304 0.055 0.122
0.688 ꢀ0.807 ꢀ0.026 ꢀ0.397
ASP
nNR2
nCaR
0.453
nCoNHR nCs
SPI
0.627
0.044
0.555
0.522
0.223
0.103
pMIC
SUMPC
Softness
Mass
SA
nBM
Me
1
0.033
1
0.755
0.697 0.727 ꢀ0.453
0.001
0.652
0.228
0.638
0.631
0.424
ꢀ0.240
ꢀ0.633
ꢀ0.413
ꢀ0.079
ꢀ0.620
ꢀ0.361
0.600
0.364
0.610
0.554
0.150
0.672
0.373
0.267 ꢀ0.072 0.473
0.30
1
0.633 0.669 ꢀ0.240 ꢀ0.153
0.604
0.325
0.675
1
0.593 ꢀ0.517
0.127
1
ꢀ0.066 ꢀ0.405
1
ꢀ0.364
0.502 ꢀ0.103 ꢀ0.372 ꢀ0.123 ꢀ0.647 ꢀ0.071 ꢀ0.081
BAC
1
ꢀ0.205
0.083
0.694 ꢀ0.227
0.420 ꢀ0.313 ꢀ0.213
0.407 ꢀ0.587
SEigz
D/Dr10
J3D
1
0.538 ꢀ0.713 ꢀ0.165 ꢀ0.596
0.327
0.673
0.488
0.188
ꢀ0.468
0.532
0.637
0.633
0.528
1
ꢀ0.447 ꢀ0.226 ꢀ0.396
ꢀ0.201
1
ꢀ0.058
0.598 ꢀ0.613 ꢀ0.440
0.640 ꢀ0.189 ꢀ0.732
0.127 ꢀ0.209 ꢀ0.330
0.346 ꢀ0.128 ꢀ0.622
ASP
1
0.097
1
0.023 ꢀ0.168
nNR2
nCaR
nCoNHR
nCs
SPI
pMIC
ꢀ0.372
0.091
0.038
1
1
ꢀ0.492 ꢀ0.011
0.585
0.341
ꢀ0.366
0.316
1
0.016 ꢀ0.556
1
0.076
1
Table 17
Correlation coefficient (r2) matrix for the descriptors of all compounds used in the different MLR equations (Pseudomonas aeruginosa).
Hardness
1
AMW
RBF
0.430
nR10
SEigz
IC3
MAXDN
nNR2
pMIC
Hardness
AMW
RBF
nR10
SEigz
IC3
MAXDN
nNR2
pMIC
ꢀ0.695
ꢀ0.282
0.241
ꢀ0.232
1
ꢀ0.739
0.739
ꢀ0.270
0.591
1
ꢀ0.277
0.183
ꢀ0.342
0.629
0.640
1
ꢀ0.705
0.597
ꢀ0.153
0.310
0.627
0.254
1
0.617
ꢀ0.555
0.618
0.050
0.378
0.622
0.102
0.464
ꢀ0.593
1
1
ꢀ0.475
ꢀ0.729
0.252
ꢀ0.445
ꢀ0.630
ꢀ0.117
ꢀ0.530
1
1
Table 18
Correlation coefficient (r2) matrix for the descriptors of compounds 9, 10, 15–24 used in the different MLR equations (Candida albicans).
HOMO
1
SA
nO
BAC
0.450
0.747
ꢀ0.478
1
PJI3
0.606
0.694
ꢀ0.760
0.768
1
G(N.N)
nCp
0.588
nCaH
pMIC
HOMO
SA
nO
BAC
PJI3
G(N.N)
nCp
nCaH
pMIC
0.629
1
ꢀ0.775
ꢀ0.823
1
0.388
0.680
ꢀ0.534
0.606
0.774
1
0.269
ꢀ0.062
0.156
ꢀ0.336
ꢀ0.085
ꢀ0.405
ꢀ0.346
1
ꢀ0.680
ꢀ0.680
0.651
ꢀ0.702
ꢀ0.739
ꢀ0.262
ꢀ0.753
ꢀ0.202
1
0.778
ꢀ0.673
0.929
0.618
0.633
1
Table 19
be noted that the chemical, constitutional, geometrical and func-
tional group descriptors did not have significant impacts on the
activity of these compounds. The first equation E1 was found by
using quantum descriptors and includes the negative effect of
HOMO on the antimicrobial activity of compounds. This contri-
bution suggests that electronic interaction plays important roles in
antimicrobial activity of these compounds. Molecules with low
HOMO energy values are more able to accept electrons than
Correlation coefficient (r2) matrix for the descriptors of compounds 9,10,15–24 used
in the different MLR equations (Pseudomonas aeruginosa).
HOMO
1
PHI
IDDE
pMIC
HOMO
PHI
IDDE
pMIC
0.626
1
0.626
0.716
1
ꢀ0.615
ꢀ0.184
ꢀ0.710
1

A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
2153
Table 20
Correlation coefficient (r2) matrix for the descriptors of compounds 25–43 used in the different MLR equations (Staphylococcus aureus).
DMy
1
nDB
BAC
nRORPh
nCs
nCaR
0.053
ꢀ0.373
ꢀ0.297
0.307
ꢀ0.378
1
nAB
0.002
RBF
VRA2
L/BW
0.164
SPAN
0.249
pMIC
DMy
nDB
BAC
nRORPh
nCs
nCaR
nAB
0.319
1
0.609
0.461
1
ꢀ0.348
ꢀ0.265
ꢀ0.677
1
ꢀ0.573
ꢀ0.084
0.028
ꢀ0.217
1
ꢀ0.235
ꢀ0.222
0.228
ꢀ0.440
0.607
ꢀ0.606
ꢀ0.670
1
0.116
0.751
0.025
0.243
ꢀ0.208
0.107
0.163
ꢀ0.559
1
0.653
0.501
0.598
ꢀ0.478
ꢀ0.347
ꢀ0.370
ꢀ0.431
ꢀ0.075
0.026
ꢀ0.396
ꢀ0.414
0.509
ꢀ0.429
0.771
1
0.280
ꢀ0.156
0.352
ꢀ0.167
ꢀ0.061
0.045
ꢀ0.462
0.301
1
0.549
ꢀ0.065
0.400
ꢀ0.472
0.364
0.429
ꢀ0.792
0.781
0.627
1
RBF
VRA2
L/BW
SPAN
pMIC
ꢀ0.042
0.030
1
molecules with high HOMO energy values. As it can be seen from
equation E2, a valuable correlation is observed between topological
descriptors (mean information content) on the distance degree
quality (IDDE), kier flexibility index (PHI) and pMIC of these
compounds. The last equation (E3) was derived from all calculated
descriptors. Stepwise selection and elimination of variables
produced a two-parametric equation which is similar to what was
obtained for E2.
MLR analysis with different types of descriptors for compounds
9, 10, 15–24 was performed, but no relationship between the
structure and the antimicrobial activity against S. aureus was
obtained.
Table 13 describes that the QSAR models resulted for the anti-
microbial activity of compounds 25–43 against S. aureus by using
different sets of molecular descriptors. The equation E1 shows that
among quantum descriptors, DMy has a positive effect on the
antimicrobial activity. This contribution suggests that electronic
interaction plays an important role in inhibitory activity of these
compounds. The second equation of Table 13 was found by using
constitutional descriptors (E2). This equation explained the positive
effect of (nDB) on the antimicrobial activity of compounds 25–43.
The effect of topological descriptors on the antimicrobial activity of
these compounds has been described by equation E3 of Table 13. It
explains the positive effect of (BAC) on the activity of these
compounds. The effect of functional group descriptors on the
antimicrobial activity of the studied compounds has been described
by equation E4 of Table 13. It explained the negative effect of
number of ethers (aromatic) (nRORPh), number of total secondary
C (SP3) (Cs) and the number of substituted aromatic C (SP2) (nCaR)
on the antimicrobial activity of the compounds. The negative sign
of the coefficient of the nRORPh, Cs and nCaR proposes that
decreasing the number of the above-mentioned descriptors,
enhances the activity. Equation E5 was obtained from the all types
of calculated descriptors. Stepwise selection and elimination of
variables produced again a six-parametric QSAR equation. This
equation, which has a high statistical quality (R2 ¼ 0.93, Q2 ¼ 0.82),
demonstrates that quantum (DMy), constitutional (nAB) & (RBF),
topological (average Randic-type eigenvector-based index from
adjacency matrix (VRA2)) and geometrical (length-to-breath ratio
by WHIM (L/BW), SPAN (Span R)) parameters are major factors that
affect the antimicrobial activity of compounds 25–43. All described
parameters except DMy and SPAN have negative effect on the
activity.
The resulted equations for the antimicrobial activity of
compounds 25–33, 35–43 against C. albicans are listed in Table 14.
In this series the quantum parameter did not have a significant
impact on the antimicrobial activity. As it can be seen from equa-
tion E1 (Table 14) mass of studied molecules has a positive effect on
the antimicrobial activity. The equation obtained from the effect of
constitutional parameter on the antimicrobial activity of the
compounds 25–33, 35–43 (E2) shows, the positive effect of the
number of aromatic bonds (nAB) on the activity of these
compounds. The positive sign of the coefficient of nAB proposes
that increasing the number of aromatic bonds of the molecule
enhances the activity. The equation found by using topological
descriptors (E3) explains the positive effect of molecular multiple
path count of order 7 (piPC07), average connectivity index chi-2
(X2A) and Eigenvalue sum from van der Waals weighted distance
matrix (SEigv) on the activity of the studied compounds. This
equation, which has a high statistical quality could explain and
predict 0.81% and 0.71% of variance in pMIC data, respectively. The
equation obtained from the effect of geometrical parameter on the
antimicrobial activity of the studied compounds (E4), shows
a positive effect of maximal electrotopological positive variation
(MAXDP) on the antimicrobial activity of compounds 25–33, 35–43.
The MLR equation of Table 14 obtained from the pool of functional
group descriptors (E5) explained the positive effect of the number
of substituted aromatic C (SP2) (nCaR) on the antimicrobial activity.
The last equation (E6) was derived from all calculated descriptors.
Stepwise selection and elimination of variables produced a three-
parametric equation. It shows the positive effect of nCaR and the
negative effect of ASP and the number of secondary amines
Table 21
Correlation coefficient (r2) matrix for the descriptors of compounds 25–33, 35–43 used in the different MLR equations (Candida albicans).
Mass
1
nAB
X2A
SEigv
piPc07
MAXDP
nCaR
0.511
ASP
0.087
nNHRPH
pMIC
0.484
Mass
nAB
X2A
SEigv
piPCO7
MAXDP
nCaR
ASP
nNHRPH
pMIC
0.525
1
ꢀ0.634
ꢀ0.61
1
ꢀ0.626
ꢀ0.100
0.377
1
0.630
0.618
ꢀ0.715
ꢀ0.470
1
0.767
0.590
ꢀ0.690
ꢀ0.503
0.655
1
0.289
0.615
ꢀ0.648
ꢀ0.075
0.643
0.383
0.717
0.069
1
0.770
ꢀ0.749
ꢀ0.112
0.669
0.586
1
0.205
ꢀ0.058
ꢀ0.104
0.021
ꢀ0.019
0.114
0.730
ꢀ0.488
ꢀ0.103
0.721
0.524
0.750
ꢀ0.307
0.316
1
1

2154
A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
activities. Some of the tested compounds were moderate to good
inhibitors of bacterial growth. Compounds 9, 20 and 21 were active
against S. aureus at 16 g/ml concentration. Compound 9 was active
against S. enteritidis at the same concentration. It was also
a moderate inhibitor of B. subtilis.
S.aureus
5.5
5
y = 0.9798x + 0.0787
R2 = 0.9779
m
Quantitative relationships between molecular structure and
antimicrobial activity of a set of ten derivatives of 3-hydroxypyridin-
4-one 9,10,15–22 which we prepared and twenty-one derivatives of
3-hydroxypyridine-4-one and 3-hydroxypyran-4-one 23–43 which
have been reported before were discovered by MLR method.
According to Table 9, geometrical parameters (J3D, ASP), functional
group descriptor (nCs) and topological parameter (SPI) have
important roles in the antimicrobial activity against C. albicans. This
4.5
4
3.5
3
3
3.5
4
4.5
5
5.5
equation has
a
good statistical quality (R2 ¼ 0.81, SE ¼ 0.14,
Experimental activity
Q2 ¼ 0.73). Results for compounds 9,10,15–24 revealed that nCp and
nCaH parameters have significant impact on the antimicrobial
activity of the compounds against C. albicans. The proposed equa-
tion containing this parameter could explain and predict 0.81% and
0.61% of variance in pMIC data, respectively. In compounds 25–43,
quantum (DMy), constitutional (nAB, RBF), topological (VRA2) and
geometrical (L/BW, SPAN) parameters are major factors that affect
antimicrobial activity against S. aureus. This equation has a high
statistical quality (R2 ¼ 0.93, SE ¼ 0.19, Q2 ¼ 0.82). As it can be seen
from E3 of Table 14 topological indices (piPC07, X2A and SEigv) have
a significant role in the antimicrobial activity against C. albicans. This
C.albicans
5
4.5
4
y = 0.9127x + 0.3293
R2 = 0.8277
3.5
3
equation has
a
high statistical quality (R2 ¼ 0.81, SE ¼ 0.09,
Q2 ¼ 0.71). According to E6, geometrical parameter (ASP) and func-
tional parameter (nCaR and nNHRPh) have also significant impact
on the antimicrobial activity against C. albicans. This equation has
a good statistical quality and can explain and predict 0.82% and
0.61% of variance in pMIC data, respectively.
3
3.5
4
Experimental activity
4.5
5
p.aeruginosa
4.5
4
In summary, all of the studied compounds showed a better
QSAR model for the antimicrobial activity against C. albicans and
compounds 25–43 had the best QSAR model for anti S. aureus
activity in comparison with other tested microorganisms.
y = 0.315x + 2.5454
R2 = 0.315
4. Experimental protocols
3.5
3
4.1. Chemistry
All chemicals used for the synthesis of the compounds were
supplied by Merck or Fluka. Melting points were determined on
a Mettler capillary melting point apparatus and were uncorrected.
The IR spectra were recorded with a Perkin Elmer 1420 Ratio
Recording IR spectrometer as a KBr disc (
spectra (DMSO-d6) were recorded on a Bruker 300 MHz spec-
trometer. Chemical shifts ( ) are reported in ppm downfield from
3
3.5
4
4.5
Experimetal activity
g
, cmꢀ1). The 1H NMR
Fig. 1. Plots of the cross-validated predicted activity against the experimental activity
for the MLR model obtained against three microorganisms.
d
the internal standard tetramethylsilane (TMS). The mass spectra
were acquired with a Finnigan TSQ-70 mass spectrometer. Elec-
tron-impact ionization was performed at an ionizing energy of
70 eV. Elemental microanalyses were within ꢂ0.4% of the theo-
retical values for C, H and N. The purity of the compounds was
checked by thin layer chromatography (TLC) on silica gel plate
using chloroform and methanol.
(aromatic) (nNHRPh) on the antimicrobial activity of compounds
25–33, 35–43. This equation, which has a high statistical quality
could explain and predict 0.82% and 0.61% of variance in pMIC data,
respectively. MLR analysis with different types of descriptors for
compounds 25–43 was performed, but no relationship between the
structure and antimicrobial activity against P. aeruginosa was
obtained. Plots of the cross-validated predicted activity against the
experimental activity for the MLR model obtained against three
microorganisms are given in Fig. 1.
4.1.1. General procedure for the synthesis of 2-alkyl-3-benzyloxy-
pyridine-4-ones (5 and 6)
The benzyl-protected alcohols 5 and 6 were prepared by
following the methodology as described by Rai and co-workers from
commercially available maltol (2-methyl-3-hydroxy-pyran-4-one) 1
or ethyl maltol (2-ethyl-3-hydroxy-pyran-4-one) 2 (Scheme 1) [42].
3. Conclusions
A series of novel Mannich bases of 2-alkyl-3-hydroxy-pyridine-4-
ones, namely 2-alkyl-3-hydroxy-5-N-piperidylmethyl or N,N-dia-
lkylaminomethyl pyridine-4-ones were prepared. The synthesized
compounds along with two N-aryl-2-methyl-3-hydroxy-pyridine-4-
one derivatives and two N-alkyl-2-methyl-3-hydroxy-pyridine-4-
one compounds were screened for their antibacterial and antifungal
4.1.2. General procedure for the preparation of Mannich bases of 2-
alkyl-3-benzyloxy-pyridine-4-ones (7, 8, 11–14)
Formaldehyde solution (40%, 1.5 mL) and the proper secondary
amine (10.0 mmol) were added to a solution of 5 or 6 (10 mmol) in

A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
2155
ethanol. The reaction mixture was refluxed for 40–80 h. The solvent
was removed by rotary evaporation under vacuum to give a yellow
to reddish brown oil. The product was taken into chloroform
(50 mL) and washed with water (2 ꢃ 10). The organic layer fraction
was dried over anhydrous sodium sulphate, filtered and concen-
trated to dryness under reduced pressure to give a yellow to
reddish brown oil.
4.1.3.2. 2-Ethyl-3-hydroxy-5-N-piperidylmethyl-pyridine-4-one
(10). M.p. 270–272 ꢁC (25%); IR: (KBr Disc) /cmꢀ1: 3200–3500
(broad, NH and OH), 1635 (C]O), 1565 (C]C); 1H NMR (DMSO-
D6):
n
d
1.35 (3H, t, CH2CH3, J ¼ 6.0 Hz), 1.41–2.04 (6H, m,
piperidyl C3, C4, C5 hydrogens), 2.92 (2H, q, CH2CH3, J ¼ 6.0 Hz),
3.41–3.52 (4H, m, piperidyl C2, C6, hydrogens), 4.35 (2H, s, CH2-
piperidyl), 5.54–6.25 (1H, br, OH), 8.31 (1H, s, 6-H); MS (EI): m/
ꢄ
ꢄ
ꢄ
z ¼ 236 [Mþ ], 235 (Mþ ꢀ H), 237 (Mþ þ H, 100%).
4.1.2.1. 2-Methyl-3-benzyloxy-5-N-piperidylmethyl-pyridine-4-one
(7). Yellow oil (40%); IR: (KBr Disc)
n
/cmꢀ1: 1655 (C]O), 1545
4.1.3.3. 2-Methyl-3-hydroxy-5(N,N-dimethyl)aminomethyl-pyridine-
(C]C); 1H NMR (DMSO-D6):
d
1.41–2.05 (6H, m, piperidyl C3, C4,
4-one (15). M.p. 252–253 ꢁC (25%); IR: (KBr Disc) /cmꢀ1
n :
C5 hydrogens), 2.25 (3H, s, 2-CH3), 2.85–3.65 (4H, m, piperidyl C2,
C6, hydrogens), 4.23 (2H, s, CH2-piperidyl), 5.22 (2H, s, CH2–Ph),
7.41 (5H, s, ArH), 8.22 (1H, s, 6-H).
3100–3400 (broad, NH and OH), 1640 (C]O), 1545 (C]C); 1H
NMR (DMSO-D6): 2.30 (3H, s, 2-CH3), 2.81 [6H, s, N(CH3)2],
d
4.22 [2H, s, CH2–N(CH3)2], 4.51–5.54 (1H, br, OH), 8.05 (1H, s,
ꢄ
ꢄ
ꢄ
6-H); MS (EI): m/z ¼ 182 [Mþ ], 181 (Mþ ꢀ H), 183 (Mþ þ H,
4.1.2.2. 2-Ethyl-3-benzyloxy-5-N-piperidylmethyl-pyridine-4-one
100%).
(8). Reddish brown oil (38%); IR: (KBr Disc)
n
/cmꢀ1: 1655
(C]O), 1570 (C]C); 1H NMR (DMSO-d6):
d
1.14 (3H, t, CH2CH3,
4.1.3.4. 2-Ethyl-3-hydroxy-5(N,N-dimethyl)aminomethyl-pyridine-
J ¼ 6.0 Hz), 1.44–2.05 (6H, m, piperidyl C3, C4, C5 hydrogens),
2.36 (2H, q, CH2CH3, J ¼ 6.0 Hz), 2.84–3.52 (4H, m, piperidyl C2,
C6, hydrogens), 4.32 (2H, s, CH2-piperidyl), 5.24 (2H, s, CH2–Ph),
7.35 (5H, s, ArH), 8.36 (1H, s, 6-H).
4-one (16). M.p. 254–255 ꢁC (30%); IR: (KBr Disc) /cmꢀ1: 3100–
n
3300 (broad, NH and OH), 1635 (C]O), 1555 (C]C); 1H NMR
(DMSO-D6):
1.22 (3H, t, CH2CH3, J ¼ 6.0 Hz), 2.60–3.00 [8H, m,
d
CH2CH3, s, N(CH3)2], 4.42 [2H, s, CH2– N(CH3)2], 4.65–5.55(1H, br,
ꢄ
ꢄ
OH), 8.25 (1H, s, 6-H),; MS (EI): m/z ¼ 196 [Mþ ], 197 (Mþ þ H,
4.1.2.3. 2-Methyl-3-benzyloxy-5(N,N-dimethyl)aminomethyl-pyridine-
100%).
4-one (11). Reddish brown oil (48%); IR: (KBr Disc) n
/cmꢀ1: 1660
1
(C]O), 1550 (C]C); H NMR (DMSO-D6):
d
2.13 (3H, s, 2-CH3), 2.74
4.1.3.5. 2-Methyl-3-hydroxy-5(N,N-diethyl) aminomethyl-pyridine-
[6H, s, N(CH3)2], 4.20 [2H, s, CH2–N(CH3)2], 5.24 (2H, s, CH2–Ph), 7.43
(5H, s, ArH), 8.02 (1H, s, 6-H).
4-one (17). M.p. 254–255 ꢁC (23%); IR: (KBr Disc) /cmꢀ1: 3100–
n
3300 (broad, NH and OH), 1640 (C]O), 1555 (C]C); 1H NMR
(DMSO-D6):
d
1.3 [6H, t, J ¼ 8.0 Hz, N(CH2CH3)2], 2.5 (3H, s, 2-CH3),
4.1.2.4. 2-Ethyl-3-benzyloxy-5(N,N-dimethyl)aminomethyl-pyridine-
3.2 [4H, q, J ¼ 8.0 Hz, N(CH2CH3)2], 4.4 [2H, s, CH2–N(CH2CH3)2],
ꢄ
4-one (12). Reddish brown oil (45%); IR: (KBr Disc)
n
/cmꢀ1: 1650
6.1–7.2 (broad, -OH), 8.4 (1H, s, 6-H); MS (EI): m/z ¼ 210 [Mþ ], 209
ꢄ
ꢄ
(C]O), 1560 (C]C); 1H NMR (DMSO-D6):
d
1.00 (3H, t, CH2CH3,
(Mþ ꢀ H), 211 (Mþ þ H, 100%).
J ¼ 6.0 Hz), 2.62–3.04 [8H, m, CH2CH3, s, N(CH3)2], 4.42 [2H, s, CH2–
N(CH3)2], 5.12 (2H, s, CH2–Ph), 7.3 (5H, s, ArH), 8.22 (1H, s, 6-H).
4.1.3.6. 2-Ethyl-3-hydroxy-5(N,N-diethyl) aminomethyl-pyridine-4-
one (18). M.p. 256–257 ꢁC (23%); IR: (KBr Disc) /cmꢀ1: 3100–3400
n
4.1.2.5. 2-Methyl-3-benzyloxy-5(N,N-diethyl)aminomethyl-pyridine-
(broad, NH and OH), 1630 (C]O), 1545 (C]C); 1H NMR (DMSO-D6):
4-one (13). Yellow oil (42%); IR: (KBr Disc)
n
/cmꢀ1: 1655 (C]O),
d
1.0–1.4 [9H, m, N(CH2CH3)2, 2-CH2CH3], 2.7 (2H, q, J ¼ 6.0 Hz, 2-
1560 (C]C); 1H NMR (DMSO-D6):
d
1.31 [6H, t, J ¼ 8.0 Hz,
CH2CH3), 3.1 [4H, q, J ¼ 8.0 Hz, N(CH2CH3)2], 4.3 [2H, s, CH2–
N(CH2CH3)2], 4.9–5.6 (broad, –OH), 8.1 (1H, s, 6-H); MS (EI): m/
N(CH2CH3)2], 2.24 (3H, s, 2-CH3), 3.25 [4H, q, J ¼ 8.0 Hz,
N(CH2CH3)2], 4.41[2H, s, CH2–N(CH2CH3)2], 5.12 (2H, s, CH2–Ph),
7.42 (5H, s, ArH), 8.40 (1H, s, 6-H).
ꢄ
ꢄ
z ¼ 224 [Mþ ], 223 (Mþ ꢀ H).
4.1.4. General procedure for the synthesis of N-aryl-2-methyl-3-
hydroxy-pyridine-4-ones (19 and 20)
4.1.2.6. 2-Ethyl-3-benzyloxy-5(N,N-diethyl)aminomethyl-pyridine-
4-one (14). Reddish brown oil (38%); IR: (KBr Disc)
n
/cmꢀ1: 1650
Synthesis of N-aryl-2-methyl-3-hydroxy-pyridine-4-ones 19
and 20 was achieved via a single step synthetic pathway. Maltol (1)
(5 mmol) was refluxed with an excess of the suitable primary aryl
amines (7.5 mmol) in an acidic solution of 9.0 mL water, 0.2 mL HCl
and 1.0 mL ethanol (pH ¼ 5) for 50–60 h. After the completion of
the reaction the reaction mixture was adjusted to pH ¼ 7 using
sodium hydroxide solution (2 N) and the product were collected by
filtration. Purification was achieved by re-crystallization from hot
methanol.
(C]O), 1550 (C]C); 1H NMR (DMSO-D6):
d
0.83–1.32 [9H, m,
N(CH2–CH3)2, 2-CH2CH3], 2.31 (2H, q, J ¼ 6.0 Hz, 2-CH2CH3), 3.13
[4H, q, J ¼ 8.0 Hz, N(CH2–CH3)2], 4.31 [2H, s, CH2–N(CH2CH3)2], 5.11
(2H, s, CH2–Ph), 7.42 (5H, s, ArH), 8.11 (1H, s, 6-H).
4.1.3. General procedure for the preparation of Mannich base
derivatives of 2-alkyl-3-hydroxy-pyridine-4-ones (9, 10, 15–18)
The benzyl-protected Mannich base derivatives 7, 8, 11–14
(0.5 mmol) were dissolved in dimethylformamide (5 mL) and Pd/C
catalyst (5%) was added. The solution was stirred at room
temperature under a constant stream of hydrogen for 2 h [43]. The
reaction mixture was then filtered and the solvent was removed
under vacuum. Recrystallization from ethanol/diethylether affor-
ded white to pale yellow crystals.
4.1.4.1. 1-Phenyl-2-methyl-3-hydroxy-pyridine-4-one (19). M.p. 221–
222 ꢁC (35%); IR: (KBr Disc)
n
/cmꢀ1: 3200 (broad, OH), 1630
(C]O), 1580 (C]C); 1H NMR (DMSO-D6):
d 2.0 (s, 3H, 2-CH3), 5.7
(bs, 3-OH), 6.2 (d, 1H, 5-H, J ¼ 8.0 Hz), 7.3–7.7 (m, 6H, ArH & 6-H).
ꢄ
MS (EI): m/z ¼ 201 [Mþ ], 200 (M ꢀ H), 184 (M ꢀ OH), 124
(M ꢀ C6H5).
4.1.3.1. 2-Methyl-3-hydroxy-5-N -piperidylmethyl-pyridine-4-one
(9). M.p. 269–270 ꢁC (22%); IR: (KBr Disc)
n
/cmꢀ1: 3100–3400
4.1.4.2. 1-(3-hydroxyphenyl)-2-methyl-3-hydroxy-pyridine-4-one
(broad, NH and OH), 1635 (C]O), 1540 (C]C); 1H NMR (DMSO-
d6): 1.32–1.95 (6H, m, piperidyl C3, C4, C5 hydrogens), 2.45 (3H,
(20). M.p. 268–269 ꢁC (40%); IR: (KBr Disc) /cmꢀ1: 3100 (broad,
n
d
OH), 1630 (C]O), 1600 (C]C); 1H NMR (DMSO-d6):
d 2.0 (s, 3H,
s, 2-CH3), 2.93–3.55 (4H, m, piperidyl C2, C6, hydrogens), 4.36
(2H, s, CH2-piperidyl), 5.80–6.75 (1H, br, OH), 8.23 (1H, s, 6-H);
2-CH3), 4.5 (bs, Ar–OH), 6.3 (d, 1H, 5-H, J ¼ 8.0 Hz), 6.8–7.7 (m,
ꢄ
ꢄ
5H, ArH & 6-H). MS (EI): m/z ¼ 217 [Mþ ], 216 (Mþ ꢀ H), 200
ꢄ
ꢄ
ꢄ
ꢄ
MS (EI): m/z ¼ 222 [Mþ ], 221 (Mþ ꢀ H), 223 (Mþ þ H, 100%).
(Mþ ꢀ OH).

2156
A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
constant values and those detected were removed from the
original data matrix. The correlation of descriptors with each
others and with the activity data was determined.
ii) The input variable in MLR must not be highly correlated.
Among the collinear descriptors detected (r > 0.9) one with
the highest correlation with the activity was retained and the
rest were omitted.
4.1.5. General procedure for the synthesis of N-alkyl-2-methyl-3-
hydroxy-pyridine-4-ones (21 and 22)
The N-alkyl derivatives of maltol 21 and 22 (Scheme 3).were
prepared following the methodology as described by Harris from
commercially available maltol 1 [38].
4.2. Antimicrobial activity determination
iii) The selected descriptors from each class and the experi-
mentally antimicrobial data were analyzed by the stepwise
regression SPSS (version 12.0) software.
Muller–Hinton broth and RPMI 1640 medium buffered with
MOPS at pH 7.0 was used for bacteria and fungi respectively [41].
The inocula of microorganism (106 c.f.u mlꢀ1) were prepared from
culture and suspensions were adjusted to 0.5 McFarland standard
turbidity. The final size of inocula was 1.5 ꢃ104 for bacteria. The test
compound dissolved in dimethyl sulfoxide (DMSO) was first diluted
to the highest concentration to be tested and DMSO had no effect
on the microorganisms in the concentrations studied. Serial two-
For the development of QSAR equations, Stepwise-MLR method
was used.
In present study, MLR with stepwise selection and elimination
of variables was applied for developing QSAR models using SPSS
software (SPSS Inc., version 12.0). The resulted models were vali-
dated by leave-one out cross-validation procedure (using MATLAB
software) to check their predictability and robustness.
fold dilutions were made in concentration range from 8
m
g mlꢀ1 to
512
m
g mlꢀ1 in sterile 96 well microplates. 20 l of Alamar BlueÒ
m
reagent was added to each well. Plates were covered and sealed
with parafilm and incubated for 24 h at 37 ꢁC. The MIC was defined
as the lowest concentration, which prevented a color change from
blue to pink. Ciprofloxacin was used as standard antibacterial drug.
The same method except some modifications was used for the
antifungal studies. The final size of inocula of microorganism was
1.5 ꢃ105 for fungi. The incubation time was 48 h at 25 ꢁC. Ketoco-
nazole was used as standard antifungal agent.
Acknowledgments
This work was supported by Isfahan Pharmaceutical Sciences
Research Center. The authors wish to thank Dr. S. Hossein Hejazi
and Dr. Mohammad Jalali for technical supports in this research.
References
[1] J.W. Eaton, P. Brandt, J.R. Mahoney, J.T. Lee, Science 215 (1982) 691–693.
[2] R.L. Jones, C.M. Peterson, R.W. Grady, T. Kumbaraci, A. Cerami, Nature 267
(1977) 63–65.
4.3. QSAR analysis
4.3.1. Software and descriptor generation
[3] E.D. Weinberg, Drug Metab. Rev. 22 (1990) 531–579.
[4] E.D. Weinberg, Microbiol. Rev. 42 (1978) 45–66.
A Pentium IV personal computer (CPU at 3.06 GHz) with
windows XP operating system was used. The two-dimensional
structures of molecules were drawn using Hyperchem 7.0 software.
The final geometries were obtained with the semi-empirical AM1
method in Hyperchem program. The molecular structures were
optimized using the Polak–Ribiere algorithm until the root mean
square gradient was 0.01 kcal molꢀ1. The resulted geometry was
transferred into Dragon program package, which was developed by
Milano chemometrics and QSAR Group [44]. The z-matrix of the
structures were provided by the software and transferred to the
Gaussian 98 program [45].
Complete geometry optimization was performed taking the
most extended conformation as starting geometries. Semi-empir-
ical molecular orbital calculation (AM1) of the structures was per-
formed using Gaussian 98 program [45].
A large number of molecular descriptors were calculated using
Hyperchem, Gaussian 98 and Dragon package [44]. Some chemical
parameters including molecular volume (V), molecular surface area
(SA), hydrophobicity (log P), hydration energy (HE) and molecular
polarizability (MP) were calculated using Hyperchem Software.
Gaussian 98 was employed for calculation of different quantum
chemical descriptors including, dipole moment (DM), local charges,
[5] J.B. Neilands, J. Biol. Chem. 270 (1995) 26723–26726.
[6] V. Braun, Int. J. Med. Microbiol. 291 (2001) 67–79.
[7] E.D. Weinberg, Physiol. Rev. 64 (1984) 65–102.
[8] R.A. Finkelstein, C.V. Sciortino, M.A. McIntosh, Rev. Infect. Dis 5 (1983) S759–
S776.
[9] E.D. Weinberg, J. Nutr. Immunol 1 (1992) 41–63.
[10] E.P. Skaar, A.H. Gaspar, O. Schneewind, J. Bacteriol. 188 (2006) 1071–1080.
[11] J.A. Washington, R.N. Jones, S.D. Allen, E.H. Gerlack, F.P. Koontz, P.R. Murray,
M.A. Pfaller, M.E. Erwin, Antimicrob. Agents Chemother. 35 (1991) 1508–1511.
[12] P. Silley, J.W. Griffiths, D. Monsey, A.M. Harris, Antimicrob. Agents Chemother.
34 (1990) 1806–1808.
[13] G.A. Weinberg, Antimicrob. Agents Chemother. 38 (1994) 997–1003.
[14] B.S. van Asbeck, J.H. Marcelis, J.J.M. Marx, A. Struyvenberg, J.H. van Kats,
J. Verhoef, Eur. J. Clin. Microbiol. Infect. Dis. 2 (1983) 426–431.
[15] Z.I. Cabantchik, S. Moody-Haupt, V.R. Gordeuk, FEMS Immunol. Med. Micro-
biol. 26 (1992) 289.
[16] B.S. van Asbeck, J.H. Marcelis, J.H. van Kats, E.Y. Jaarsma, J. Verhoef, Eur. J. Clin.
Microbiol. Infec. Dis. 2 (2005) 432–438.
[17] M. Veverka, Chem. Pap-Chem. Zvesti. 46 (1992) 206–210.
[18] D.D. Erol, N. Yulug, Eur. J. Med. Chem. 29 (1994) 893–897.
[19] F. Min-Hua, L. van der Does, A. Bantjes, J. Med. Chem. 36 (1993) 2822–2827.
[20] S. Masoud, S.A. El-Tahana, A. El-Enein, Transit. Met. Chem. 14 (1989) 155–156.
[21] R. Petrola, L. Repetti, Finn. Chem. Lett. 5 (1985) 213–218.
[22] T. Kotani, I. Ichimoto, C. Tatsumi, T. Fujita, Agr. Biol. Chem. Tokyo 39 (1975)
1311–1317.
[23] T. Kotani, I. Ichimoto, C. Tatsumi, T. Fujita, Agr. Biol. Chem. Tokyo 40 (1976)
765–770.
[24] K. Iwamatsu, K. Atsumi, K. Sakagami, H. Ogino, T. Yoshida, T. Tsuruoka,
S. Shibahara, S. Inouye, S. Kondo, J. Antibiot. 43 (1990) 1450–1463.
[25] H. Ogino, K. Iwamatsu, K. Katano, S. Nakabayashi, T. Yoshida, T. Tsuruoka,
S. Inouye, S. Kondo, J. Antibiot. 43 (1990) 174–188.
HOMO and LUMO energies, hardness (h), softness (S), electroneg-
ativity ( ) and electrophilicity ( ) according to the method
c
u
proposed by Thanikaivelan et al. [46]. Dragon software calculated
different functional groups, topological, geometrical and constitu-
tional descriptors for each molecule.
[26] M.D. Aytemir, D.D. Erol, R.C. Hider, M. Ozalp, Turk. J. Chem. 27 (2003) 757–764.
[27] M.D. Aytemir, R.C. Hider, D.D. Erol, M. Ozalp, M. Ekizoglu, Turk. J. Chem. 27
(2003) 445–452.
[28] C. Hansch, D. Hoekman, H. Gao, Chem. Rev. 96 (1996) 1045–1075.
[29] H. Schmidi, Chemom. Intell. Lab. Syst. 37 (1997) 125–136.
[30] C.A. Hansch, A. Kurup, R. Garg, H. Gao, Chem. Rev. 101 (2001) 619–672.
[31] S. Wold, J. Trygg, A. Berglund, H. Antii, Chemom. Intell. Lab. Syst. 58 (2001)
131–150.
[32] B. Hemmateenejad, R. Miri, M. Akhond, M. Shamsipur, Chemom. Intell. Lab.
Syst. 64 (2002) 91–99.
[33] B. Hemmateenejad, R. Miri, M. Akhond, M. Shamsipur, Arch. Pharm. Pharm.
Med. Chem. 10 (2002) 472–480.
4.3.2. Data processing and modeling
The selection of significant descriptors, which related the anti-
microbial data to the molecular structures, is an important step in
QSAR modeling. Selection of significant descriptors was performed
through the following steps:
[34] D. Horvath, B. Mao, QSAR Comb. Sci. 22 (2003) 498–509.
[35] S. Putta, J. Eksterowicz, C. Lemmen, R. Stanton, J. Chem. Inf. Comput. Sci. 43
(2003) 1623–1635.
i) The calculated descriptors were collected in a data matrix, D.
First the descriptors were checked for constant or near

A. Fassihi et al. / European Journal of Medicinal Chemistry 44 (2009) 2145–2157
2157
[36] S. Gupta, M. Singh, A.K. Madan, J. Chem. Inf. Comput. Sci. 39 (1999) 272–
277.
[42] B.L. Rai, L.S. Dehkordi, H. Khodr, Y. Jin, Z. Liu, R.C. Hider, J. Med. Chem. 41
(1998) 3347–3359.
[37] V. Consonni, R. Todeschini, M. Pavan, J. Chem. Inf. Comput. Sci. 42 (2002) 693–
705.
[43] B.L. Rai, Z.D. Liu, D.Y. Liu, S.L. Lu, R.C. Hider, Eur. J. Med. Chem. 34 (1999) 475.
[44] R. Todeschini, Milano Chemometrics and QSPR Group, http://www.disat.
unimib.it/vhml.
[45] M.J. Frisch, M.J. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
V.G. Zakrzewski, et al., Gaussian 98, Revision A.7, Gaussian Inc, Pittsburgh, PA, 1998.
[46] P. Thanikaivelan, V. Subramanian, J.R. Rao, B.U. Nair, Chem. Phys. Lett. 323
(2000) 59–70.
[38] R.L.N. Harris, Aust. J. Chem. 29 (1976) 1329–1334.
[39] C.N. Baker, F.C. Tenover, J. Clin. Microbiol. 34 (1996) 2654–2659.
[40] W.-K. To, A.W. Fothergill, M.G. Rinaldi, J. Clin. Microbiol. 33 (1995) 2660–2664.
[41] C. Kirilmis, M. Ahmedzade, S. Servi, M. Koca, A. Kizirgil, C. Kazaz, Eur. J. Med.
Chem. 43 (2008) 300–308.