Synthesis, antibacterial activity, and quantitative structure-activity relationships of new (Z)-2-(nitroimidazolylmethylene)-3(2 H)-benzofuranone derivatives

Article abstract of DOI:10.1016/j.bmcl.2007.09.062

A new series of (Z)-2-(1-methyl-5-nitroimidazole-2-ylmethylene)-3(2 H)-benzofuranones (11a-p) and (Z)-2-(1-methyl-4-nitroimidazole-5-ylmethylene)-3(2 H)-benzofuranones (12a-m) were synthesized and assayed for their antibacterial activity against Gram-positive and Gram-negative bacteria. Most of the 5-nitroimidazole analogues (11a-p) showed a remarkable inhibition of a wide spectrum of Gram-positive bacteria (Staphylococcus aureus, Streptococcus epidermidis, MRSA, and Bacillus subtilis) and Gram-negative Klebsiella pneumoniae, whereas 4-nitroimidazole analogues (12a-m) were not effective against selected bacteria. The quantitative structure-activity relationship investigations were applied to find out the correlation between the experimentally evaluated activities with various parameters of the compounds studied. The QSAR models built in this work had reasonable predictive power and could be explained by the observed trends in activities.

Full text of DOI:10.1016/j.bmcl.2007.09.062

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 17 (2007) 6354–6363
Synthesis, antibacterial activity, and quantitative structure–activity
relationships of new (Z)-2-(nitroimidazolylmethylene)-
3(2H)-benzofuranone derivatives
Narges Hadj-esfandiari,a Latifeh Navidpour,a Hooman Shadnia,b Mohsen Amini,a
Nasrin Samadi,c Mohammad Ali Faramarzid and Abbas Shafieea,*
aDepartment of Medicinal Chemistry, Faculty of Pharmacy and Pharmaceutical Sciences Research Center,
Tehran University of Medical Sciences, Tehran 14174, Iran
bDepartment of Chemistry, Carleton University, Ottawa, Ont., Canada K1S 5B6
cDepartment of Drug and Food Control, Faculty of Pharmacy and Pharmaceutical Sciences Research Center,
Tehran University of Medical Sciences, Tehran 14174, Iran
dDepartment of Pharmaceutical Biotechnology, Faculty of Pharmacy, Tehran University of Medical Sciences, Tehran 14174, Iran
Received 14 January 2007; revised 2 August 2007; accepted 18 September 2007
Available online 21 September 2007
Abstract—A new series of (Z)-2-(1-methyl-5-nitroimidazole-2-ylmethylene)-3(2H)-benzofuranones (11a–p) and (Z)-2-(1-methyl-4-
nitroimidazole-5-ylmethylene)-3(2H)-benzofuranones (12a–m) were synthesized and assayed for their antibacterial activity against
Gram-positive and Gram-negative bacteria. Most of the 5-nitroimidazole analogues (11a–p) showed a remarkable inhibition of a
wide spectrum of Gram-positive bacteria (Staphylococcus aureus, Streptococcus epidermidis, MRSA, and Bacillus subtilis) and
Gram-negative Klebsiella pneumoniae, whereas 4-nitroimidazole analogues (12a–m) were not effective against selected bacteria.
The quantitative structure–activity relationship investigations were applied to find out the correlation between the experimentally
evaluated activities with various parameters of the compounds studied. The QSAR models built in this work had reasonable pre-
dictive power and could be explained by the observed trends in activities.
Ó 2007 Elsevier Ltd. All rights reserved.
The emergence of bacterial resistance to established
antibiotics in community and hospital-acquired infec-
tions is an ever-growing concern. In particular, the big
three Gram-positive species, Staphylococcus, Enterococ-
cus, and Streptococcus, have all evolved resistant strains
over the last few years: methicillin-resistant Staphylococ-
cus aureus (MRSA) strains now account for over 40% of
S. aureus infections in hospitals;1 vancomycin-resistant
entrococcal (VRE) infections have also increased sub-
stantially, with some strains resistant to all established
antibiotics2; Streptococcus pneumoniae is the most com-
mon cause of community-acquired pneumoniae and is
increasingly resistant to b-lactams and macrolids. For
these reasons, there is an urgent need for new antibacte-
rial agents.
The use of nitroheterocyclic drugs as antibacterial, anti-
protozoal, and anti-cancer agents is well established.
5-Nitroimidazoles and 5-nitrofuranes are the classes of
nitroheterocyclic drugs most used as antibacterials.3
The distinguishing properties between these two sub-
classes are their reduction potentials. It has been pro-
posed that the active damaging species, intermediates
in the reduction process, is the one-electron nitro radical
anion, which explains the relative cytotoxicity of these
drugs under hypoxia.4 This intermediate species inter-
acts with DNA, causing strand breaks5 with simulta-
neous release of thymine and thymidine phosphonates
and helix destabilization. Although there is a large
amount of experimental work on nitroimidazoles, they
remain an area of active research interest.
Recently, a novel class of (Z)-2-(5-nitrofuran-2-yl-meth-
ylene)-3(2H)-benzofuranones 1 has been designed and
exhibited very good in vitro antibacterial properties
(Fig. 1).6 In this regard, the QSAR analyses suggested
that the benzofuranic moiety is an important structural
Keywords: Nitroimidazole; 3(2H)-Benzofuranone; Antibacterial;
QSAR.
*
Corresponding author. Tel.: +98 21 66406757; fax: +98 21
66461178; e-mail: ashafiee@ams.ac.ir
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.09.062

N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
6355
O
methylene)-3(2H)-benzofuranones (11a–p), substituted
(Z)-2-(1-methyl-4-nitroimidazole-5-ylmethylene)-3(2H)-
benzofuranones (12a–m), (E)-2-(1-methyl-5-nitroimi-
dazole-2-ylmethylene)-1-indanone (15), and indandione
derivative (16) are outlined in Schemes 1–4.
O
R
R
O
O
HO
O
NO2
OH
2
3(2H)-Benzofuranones 8 are key intermediates for the
production of the desired compounds (11 and 12). From
our experience, the route shown in Scheme 1 summa-
rizes the most convenient synthetic path leading to the
3(2H)-benzofuranones 8. The latter was prepared start-
ing from appropriate salicylic acid derivative 3. Esterifi-
cation of 3 with ethanol in the presence of sulfuric acid
gave the ethyl salicylate 4. Treatment of ethyl ester 4
with ethyl bromoacetate generated 5,14 which were
hydrolyzed in basic media to produce 6.15 Ring closure
of 6 in refluxing AcOH=Ac2O in the presence of sodium
acetate followed by hydrolyzing acetyl ester in acidic
media [HCl/H2O/MeOH (1:10:40)] afforded the required
3(2H)-benzofuranones 8.16
1
Figure 1. Representative examples of benzofuranone antimicrobials.
feature contributing to the activity. Furthermore, a ser-
ies of (Z)-2-benzylidene-6,7-dihydroxy-3(2H)-benzof-
uranones 2 have shown antibacterial activity by
inhibiting the chorismate synthase, a key enzyme in
the shikimic acid pathway which is essential for the syn-
thesis of aromatic amino acids in bacteria.7 Considering
the above results and as part of our ongoing program in
developing new active antimicrobials against resistant
bacteria,8–13 we describe herein the synthesis and
in vitro antimicrobial properties of a novel group of
(Z)-2-(1-methyl-5-nitroimidazole-2-ylmethylene)-3(2H)-
benzofuranones (11a–p) and (Z)-2-(1-methyl-4-nitroimi-
dazole-5-ylmethylene)-3(2H)-benzofuranones (12a–m).
To gain an insight into physicochemical and the struc-
tural features of this class of compounds that are impor-
tant to the antibacterial activity, QSAR analyses of title
compounds (11 and 12) were also performed.
Title compounds, namely substituted (Z)-2-(1-methyl-
5-nitroimidazole-2-ylmethylene)-3(2H)-benzofuranones
(11a–l and 11n) as well as indanone and indandione
derivatives (15–16), were prepared by the condensation
of appropriate 3(2H)-benzofuranone 8, 1-indanone 13
or 1,3-indandione 14, respectively, with 1-methyl-5-ni-
tro-2-carbaldehyde in acetic acid, in the presence of cat-
alytic amount of sulfuric acid, while hydroxyl
substituted (Z)-2-(1-methyl-5-nitroimidazole-2-ylmeth-
Chemistry. The synthetic reactions used for the synthesis
of substituted (Z)-2-(1-methyl-5-nitroimidazole-2-yl-
COOH
COOEt
COOEt
OH
a
b
R
R
R
OH
OCH2COOEt
5
4
3
OAc
O
COOH
e
c
d
R
R
R
O
O
OCH2COOH
8
7
6
Scheme 1. Reagents and conditions: (a) EtOH (abs), H2SO4 (catalytic), reflux, 48 h; (b) BrCH2COOEt, NaI, K2CO3, DMF, rt, 24 h; (c) KOH/MeOH
(10%), reflux, 3 h; (d) sodium acetate, acetic acid/acetic anhydride, reflux, 4 h; (e) HCl/H2O/MeOH (1:10:40), reflux, 1 h.
O
N
CH3
R
a or b
O2N
CHO
N
N
O
NO2
CH3
N
O
a, 11a-l. n
b, 11o-p
9
11a-l, n-p
or
+
R
O
NO2
N
N
O
8
b
CH3
R
CHO
N
O
H3C
10
O2N
N
12a-l
Scheme 2. Reagents and conditions: (a) acetic acid/sulfuric acid, 100 °C, 6 h; (b) acetic anhydride/anhydrous sodium acetate, 100 °C, 90 min.

6356
N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
O
O
O2N
CH3
CH3
a
N
N
O
O
NO2
NO2
N
N
11m
11a
O
O
O2N
CH3
CH3
a
N
O
N
O
O2N
12m
N
O2N
12a
N
Scheme 3. Reagents and condition: (a) HNO3, H2SO4, rt, 1 h.
O
O
CH3
a
N
NO2
N
N
13
15
O2N
CHO
N
+
or
CH3
O
O
O
9
a
CH3
N
NO2
O
N
14
16
Scheme 4. Reagents and condition: (a) acetic acid/sulfuric acid, 100 °C, 6 h.
ylene)-3(2H)-benzofuranones (11o–p) and substituted
(Z)-2-(1-methyl-4-nitroimidazole-5-ylmethylene)-3(2H)-
benzofuranones (12a–l) were prepared by a different
method in acetic anhydride in the presence of anhydrous
sodium acetate (Scheme 2).17,18 The nitro derivatives
(11m and 12m) were obtained by the treatment of the
unsubstituted derivatives (11a and 12a) with concen-
trated nitric acid in sulfuric acid (Scheme 3).6
systematic conformational search was then performed
(all rotatable bonds were rotated by 30 degree steps,
RMS gradient = 0.01 (default), and RMS distance was
reduced to 0.02). Generated conformers were optimized
using MMFF94s21,22 with born solvation model23 and
lowest energy conformer for each ligand was selected
and used for QSAR analysis.
Dataset: To build the QSAR models, the logarithm of
activities (1/MIC) was calculated. MIC = 100 lg/mL
Condensation gave single isomer as shown by NMR
experiment. Precedent literature19 suggested that Z-iso-
mer would be obtained. The physicochemical data of
the synthesized compounds (11a–p, 12a–m, 15, and 16)
are summarized in Table 1.
was used for the compounds which had
a
MIC > 100 lg/mL. The QuaSAR module of MOE was
used to calculate molecular descriptors. These include
the two-dimensional descriptors, and 3D descriptors as
classified into i3D (internal coordinate dependent) and
x3D (internal coordinate independent) descriptors.
5-Nitrofuryl and phenyl derivatives of substituted
3(2H)-benzofuranones (1–2, Fig. 1) were prepared as re-
ported previously.6,7
The following descriptors were calculated: Kier topolog-
ical descriptors defined by Kier and Hall (Kier1, Kier2,
Kier3, KierA1, KierA2, KierA3, and Kierflex),24 Molar
refractivity (mr), molecular refractivity (including impli-
cit hydrogens) using atomic contribution model (smr)25
van der Walls surface area and its fractions (total:
VSA, hydrogen bond acceptor: vsa_acc; hydrophobic:
vsa_hyd, and polar: vsa_pol), principle moment of iner-
tia (pmi), water accessible surface area and its fractions
(ASA, ASA+, ASAꢀ, ASA_H, ASA_P, CASA+,
CASAꢀ, FASA+, FASAꢀ, FASA_H, FASA_P, FCA-
QSAR analysis. Even though these compounds do not
build an ideal dataset for advanced QSAR studies (dis-
cussion follows), a simple QSAR study reveals satisfac-
tory findings in terms of predictability of activities.
Modeling and calculations were done on a 3.00 GHz
Pentium 4 Machine.
Modeling: Structures of compounds 11a–p and 12a–m
were sketched using builder module of MOE 2006.20
A

N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
6357
Table 1. Physicochemical data of 11a–p, 12a–m, 15, and 16
Running a non-biased fit on such a dataset leads to
models that simply discriminate between the two clas-
ses of molecules (11 vs 12), especially when using sev-
eral connectivity/geometrical descriptors. To avoid
that, the data on active molecules were assigned a high-
er ‘importance weight’ (10) versus the inactive com-
pounds (assigned weight = 1). Also, recruiting more
than three descriptors was not allowed, which in turn
limited the expected quality for best fits. Also using
more than one geometrical/topological descriptor was
not allowed.
O
O
CH3
R
R
N
O
O
CH3
N
NO2
N
O2N
N
11a-p
12a-m
Compound
R
Mp (°C)
Yield
(%)
Molecular
formula
Using the QuaSAR-Model in MOE, several PLS models
were built in accordance with the remarked consider-
ations. Performance criteria (RMSE: root mean square
error and r2: correlation coefficient for PLS models)
were used to identify best models. The best models on
each target organism are described in Table 3. Table 4
shows the predicted log activities using leave one out
cross validation method (LOO), along with residual
and Z-score values. Compounds having a Z-score of
2.5 or higher which are considered to be outliers to the
fitted model and were excluded from regression models
are shown. Figure 2a–e show the performance of these
models as plot of experimental versus predicted log
activities.
11a
11b
11c
11d
11e
11f
11g
11h
11i
H
239–241
276–278
254–257
251–253
268–270
283–285
266–268
263–266
223–225
255–257
229–231
235–237
244–247
272–274
242–244
248–249
261–264
273–276
267–269
256–257
286–288
242–244
283–285
277–278
254–257
287–289
266–268
265–267
255–257
270–272
199–201
56
54
62
61
59
58
46
54
56
42
55
48
45
57
52
47
60
46
45
65
50
62
26
43
34
60
52
41
39
56
67
C13H9N3O4
5-Cl
5-Br
5-Me
C13H8ClN3O4
C13H8BrN3O4
C14H11N3O4
C14H11N3O5
C13H8IN3O4
C13H8ClN3O4
C14H11N3O4
C14H11N3O5
C13H8ClN3O4
C14H11N3O4
C14H11N3O5
C13H8N4O6
C15H13N3O6
C13H9N3O5
C13H9N3O6
C13H9N3O4
C13H8ClN3O4
C13H8BrN3O4
C14H11N3O4
C14H11N3O5
C13H8IN3O4
C13H8ClN3O4
C14H11N3O4
C14H11N3O5
C13H8ClN3O4
C14H11N3O4
C14H11N3O5
C13H8N4O6
C14H11N3O3
C14H9N3O4
5-OMe
5-I
6-Cl
6-Me
6-OMe
7-Cl
11j
11k
11l
7-Me
7-OMe
5-NO2
6,7-(OMe)2
6-OH
4,6-(OH)2
H
11m
11n
11o
11p
12a
12b
12c
12d
12e
12f
12g
12h
12i
Lack of numerous ‘moderately active’ compounds
makes it easy to build binary models in which the com-
pounds are classified into active/inactive.27 As expected,
generally the same set of descriptors used to build PLS
models was found to be proper for binary models. A
binary threshold of 1.5 was used for logarithm of activ-
ities. Table 5 and Figure 3 show the details and perfor-
mance of binary models. Performance of binary models
is displayed as percentage of overall correct predictions
(a) and its significance or p-value (s) which is interpreted
as the likelihood of observing a result ‘at least this ex-
treme,’ under the assumption that the model is doing
no better than pure chance. The smaller the p-value,
the less likely that the total accuracy is simply a chance.
Also percentages of correctly predicted active/inactive
molecules (a1,a0) and number of active versus inactive
molecules (A/I) are given.
5-Cl
5-Br
5-Me
5-OMe
5-I
6-Cl
6-Me
6-OMe
7-Cl
12j
12k
12l
7-Me
7-OMe
5-NO2
—
12m
15
16
—
SA+, and FCASAꢀ), polar surface area calculated
using group contributions (tpsa),26 log of octanol/water
partition coefficient (logP), globularity (glob), and van
der Waals volume calculated using 0.75 A grid (vol).
For brevity, only the descriptors which were used in best
models are shown in Table 2.
Results and discussion. The antibacterial properties of
novel nitroimidazole derivatives were performed
in vitro by a conventional agar-dilution method against
a panel of selected Gram-positive and Gram-negative
bacteria (Table 6).28,29 The minimal inhibitory concen-
tration (MIC, lg/mL) was determined in comparison
with amoxicillin as the reference drug.
˚
Data analysis method: An ideal dataset for a QSAR
study should include numerous structures of diverse nat-
ure with proper span of activities. Our dataset included
two classes of molecules, one of which was almost com-
pletely inactive. On average only 12 compounds were ac-
tive on each organism. The activity span was not ideal
either; about 65% inactive, 30% highly active, and only
5% moderately active (10 lg/mL < MIC < 99 lg/mL).
Also, activity of all the inactive compounds was set to
100, an approximation that causes difficulty for fitting
models. Despite these conditions, as it follows, a care-
fully designed basic QSAR reveals interesting findings,
and practical predictability.
Generally, most 5-nitroimidazole analogues (11a–m) re-
vealed respectable in vitro activity against Gram-posi-
tive and Gram-negative bacteria, while none of the
compounds having 4-nitroimidazole moiety (12a–m) or
phenyl group (2) exhibited any antibacterial activity up
to the concentration of 100 lg/mL.
Having different cell wall structure, Klebsiella pneumo-
niae, a Gram-negative bacterium exhibited better sus-
ceptibility than Gram-positive bacteria, Staphylococci,

6358
N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
Table 2. Calculated descriptors which were used in derived models
Compound
Descriptor
KierA3
Kier1
Kier2
Kier3
KierA2
Dipole_X
vsa_pol
ASA_P
CASAꢀ
2
13.959
14.917
15.879
15.879
15.879
16.844
15.879
15.879
15.879
16.844
15.879
15.879
16.844
17.811
18.781
15.879
16.844
14.917
15.879
15.879
15.879
16.844
15.879
15.879
16.844
15.879
15.879
16.844
17.811
14.917
15.879
5.780
6.012
6.246
6.246
6.246
6.857
6.246
6.246
6.246
6.857
6.246
6.246
6.857
7.087
7.709
6.246
6.481
6.012
6.246
6.246
6.246
6.857
6.246
6.246
6.857
6.246
6.246
6.857
7.087
6.012
6.246
2.741
2.845
3.062
3.062
3.062
3.299
3.062
3.062
3.062
3.299
2.933
2.933
3.165
3.520
3.486
3.062
3.165
2.845
3.062
3.062
3.062
3.299
3.062
3.062
3.299
2.933
2.933
3.165
3.520
2.845
2.813
3.907
3.938
4.536
4.794
4.188
4.695
5.025
4.536
4.188
4.695
4.536
4.188
4.695
4.683
5.614
4.316
4.566
4.065
4.670
4.934
4.316
4.830
4.670
4.316
4.830
4.670
4.316
4.830
4.813
4.065
4.092
1.820
1.832
2.196
2.326
2.020
2.228
2.443
2.196
2.020
2.228
2.103
1.935
2.138
2.291
2.512
2.085
2.202
1.894
2.264
2.397
2.085
2.294
2.264
2.085
2.294
2.168
1.997
2.202
2.357
1.894
1.811
0.199
ꢀ0.559
ꢀ0.147
ꢀ0.240
ꢀ0.660
ꢀ0.562
ꢀ0.232
ꢀ0.325
ꢀ0.640
ꢀ0.406
ꢀ0.775
ꢀ0.597
ꢀ0.970
0.616
43.205
21.753
21.753
21.753
21.753
24.257
21.753
21.753
21.753
24.257
21.753
21.753
24.257
21.753
26.761
35.320
48.887
21.753
21.753
21.753
21.753
24.257
21.753
21.753
24.257
21.753
21.753
24.257
21.753
19.249
32.816
126.620
142.901
143.269
143.066
143.608
182.955
141.771
143.265
143.610
179.016
144.341
142.567
180.622
225.217
211.055
203.545
238.637
168.700
169.555
169.500
169.278
208.310
169.530
170.246
209.360
164.640
162.861
200.718
249.183
149.941
181.684
617.920
672.267
816.828
863.307
604.296
657.101
869.485
821.169
611.035
652.113
807.527
597.650
650.245
1057.487
617.069
762.599
847.682
677.171
839.293
889.278
621.771
662.750
828.592
620.152
671.420
805.846
598.414
644.084
1072.686
680.556
793.675
11a
11b
11c
11d
11e
11f
11g
11h
11i
11j
11k
11l
11m
11n
11o
11p
12a
12b
12c
12d
12e
12g
12h
12i
ꢀ0.245
ꢀ0.554
ꢀ0.065
0.541
0.264
0.322
0.579
0.789
0.479
0.835
0.808
0.784
12j
12k
12l
0.819
1.252
12m
15
0.039
ꢀ0.528
ꢀ0.809
16
Table 3. Details of best PLS models (n = 31)
Organism
S. aureus
Model
r
r2
RMSE
Fa
q2b
Oc
2
loga = ꢀ4.95 + 4.43 Kier1 ꢀ 2.47 Dipole_X ꢀ 0.01 CASAꢀ
loga = ꢀ4.95 + 4.43 Kier1 ꢀ 2.47 Dipole_X ꢀ 0.01 CASAꢀ
loga = ꢀ6.37 + 0.49 Kier3 ꢀ 0.51 Dipole_X ꢀ 0.49 CASAꢀ
loga = ꢀ5.52 + 4.7 Kier3 ꢀ 2.50 Dipole_X ꢀ 0.1 CASAꢀ
0.84
0.85
0.88
0.85
0.88
0.76
0.79
0.78
0.72
0.77
0.88
0.89
0.83
1.09
1.13
87.77
69.53
94.40
98.59
62.94
0.62
0.72
0.72
0.77
0.70
*
*
*
S. epidermidis
MRSA
2
*
*
*
2
*
*
*
B. subtilis
2
*
*
*
K. pneumoniae
a Fischer statistic.
loga = ꢀ14.17 + 7.11 Kier 3 ꢀ 2.44 Dipole_X ꢀ 0.07 CASAꢀ
2
*
*
*
b Cross validated square correlation coefficient.
c Number of outliers (determined using Z-scores of cross validation test, see Table 4).
and Bacillus subtilis to the action of 5-nitroimidazole
analogues (11a–p). These compounds were inactive
against Pseudomonas aeruginosa and Escherichia coli.30
(Z)-2-(1-Methyl-5-nitroimidazole-2-ylmethylene)-3(2H)-
benzofuranone (11a) was slightly superior in inhibiting
the growth of K. pneumoniae (2- to 8-fold) than the
Gram-positive bacteria.
compound with a different structural backbone is
unknown.
While Dipole_X is a good predictor, neither total dipole
nor the other components of dipole vector (Y and Z)
have comparable predictive power. Further analysis on
compounds 11a–p only (not shown here) suggests that
this predictor does more than simple discrimination of
compounds 11a–p from compounds 12a–m, so it is likely
that the negative charge of –NO2 group on C-5 of imid-
azole is necessary for action. Substitution of 5-nitroimi-
dazole with 5-nitrofurane resulted in the compound with
a slightly better inhibitory activity against Staphylococci
Since the activity profiles of the compounds on target
organisms are very similar (correlation coefficients
r2 ¼ 0:78–0:99), it is understandable that the best QSAR
models are very similar. Despite apparent robustness of
these models, their accuracy for prediction of a new

N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
Table 4. Calculated log activities with residual and Z-score values for best PLS models
6359
Structure
S. aureus
S. epidermidis
MRSA
clog(a)
Residual
Z-Score
clog(a)
Residual
Z-Score
clog(a)
Residual
Z-Score
2
1.215
2.683
—
ꢀ0.282
1.780
—
0.312
2.187
—
1.302
2.876
—
ꢀ0.369
1.588
—
0.403
1.871
—
1.097
2.754
—
ꢀ0.164
1.016
—
0.191
1.225
—
11a
11b
11c
11d
11e
11f
11g
11h
11i
1.399
4.803
5.197
1.278
2.075
4.673
4.648
2.203
4.015
5.469
—
ꢀ0.146
ꢀ0.289
ꢀ0.628
0.101
0.161
0.319
0.702
0.111
1.089
0.175
0.687
3.208
0.555
0.542
—
1.619
5.210
5.765
1.497
2.314
4.968
5.195
2.419
4.195
6.008
—
ꢀ0.366
ꢀ0.697
ꢀ1.197
ꢀ0.118
ꢀ1.197
0.240
0.400
0.768
1.358
0.128
1.362
0.262
0.074
2.742
1.136
0.056
—
1.361
4.836
5.283
1.238
2.051
4.603
4.730
2.156
4.014
5.533
—
ꢀ0.108
ꢀ0.322
ꢀ0.714
0.141
ꢀ0.934
0.606
0.534
2.428
0.500
0.424
—
0.125
0.376
0.844
0.164
1.120
0.714
0.626
3.584
0.586
0.496
—
ꢀ0.958
ꢀ0.159
0.616
0.068
2.164
1.014
11j
2.380
0.499
11k
11l
0.487
—
ꢀ0.052
—
1.085
11m
11n
11o
11p
11a
12b
12c
12d
12e
12g
12h
12i
5.586
3.057
1.640
0.125
0.323
ꢀ0.311
1.564
1.736
ꢀ0.137
0.911
1.601
ꢀ1.360
0.550
0.009
0.764
3.062
2.695
ꢀ0.227
0.077
0.251
0.085
0.591
0.991
0.902
1.786
0.585
0.718
1.428
0.155
0.565
2.865
0.565
1.243
0.439
2.663
1.978
5.660
3.283
1.770
0.181
0.470
ꢀ0.194
1.795
2.094
ꢀ0.004
1.121
1.952
ꢀ1.335
0.683
0.245
1.160
3.245
2.623
1.210
0.597
0.035
0.913
0.723
1.625
0.835
1.109
1.257
0.081
0.950
2.790
0.407
0.959
0.010
2.934
0.987
5.774
3.052
1.630
0.061
0.284
ꢀ0.361
1.564
1.780
ꢀ0.178
0.905
1.643
ꢀ1.436
0.517
0.014
0.794
3.017
2.613
ꢀ0.416
0.082
ꢀ0.521
0.937
0.833
1.614
ꢀ0.515
ꢀ0.678
1.295
0.143
ꢀ0.540
2.554
0.532
1.089
0.357
ꢀ2.026
ꢀ1.572
0.485
0.096
0.611
1.121
0.996
1.944
0.615
0.809
1.555
0.171
0.645
3.122
0.635
1.304
0.426
2.775
1.980
0.544
0.032
ꢀ0.531
0.873
0.795
0.817
0.647
1.447
1.564
ꢀ0.516
ꢀ0.633
1.254
ꢀ0.747
ꢀ0.991
1.122
0.137
ꢀ0.073
ꢀ0.850
2.453
ꢀ0.499
2.478
0.498
12j
12k
12l
0.365
0.858
1.094
0.387
12m
15
ꢀ0.009
ꢀ2.255
ꢀ0.889
ꢀ2.072
ꢀ1.654
16
Structure
B. subtilis
K. pneumoniae
clog(a)
Residual
Z-Score
clog(a)
Residual
Z-Score
2
0.430
—
—
0.504
—
—
0.452
—
—
ꢀ0.373
5.363
4.811
5.231
6.472
6.557
5.414
5.416
6.405
6.013
6.647
5.831
7.821
3.867
5.257
4.522
2.361
2.215
3.693
3.514
2.732
2.349
3.039
1.932
2.412
1.700
1.593
0.559
1.306
1.182
1.160
1.145
1.045
1.027
0.444
0.106
0.080
3.800
0.481
0.164
0.553
1.222
0.666
1.707
0.049
2.003
0.001
1.095
1.078
2.307
2.018
1.496
1.355
1.710
0.782
1.160
0.515
0.481
0.481
11a
11b
11c
11d
11e
11f
11g
11h
11i
2.501
5.122
6.271
2.421
2.611
4.925
5.854
2.777
3.934
6.402
0.811
7.219
3.799
2.948
0.114
1.285
0.791
2.034
2.982
0.822
1.396
2.870
ꢀ0.778
0.609
0.905
ꢀ1.248
0.086
ꢀ0.314
ꢀ1.042
2.667
0.977
0.103
1.806
1.275
ꢀ0.446
2.419
ꢀ0.475
0.029
ꢀ1.839
0.884
ꢀ0.168
0.462
ꢀ0.985
ꢀ1.879
0.295
ꢀ0.348
ꢀ1.767
1.895
0.439
0.197
1.153
0.077
0.282
0.953
2.855
0.891
0.092
1.724
1.179
0.400
2.302
0.426
0.026
1.774
0.812
0.154
0.424
0.906
1.738
0.271
0.319
1.632
1.752
0.403
0.181
ꢀ0.512
0.123
0.093
ꢀ3.342
0.555
0.190
0.636
ꢀ1.369
0.764
11j
11k
11l
ꢀ1.864
0.057
2.156
11m
11n
11o
11p
12a
12b
12c
12d
12e
12g
12h
12i
0.001
ꢀ1.252
ꢀ1.217
ꢀ2.576
ꢀ2.261
ꢀ1.684
1.526
ꢀ1.921
ꢀ0.884
ꢀ1.309
ꢀ0.583
ꢀ0.544
0.544
12j
12k
12l
(continued on next page)

6360
N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
Table 4 (continued)
Structure
B. subtilis
K. pneumoniae
clog(a)
Residual
Z-Score
clog(a)
Residual
Z-Score
12m
15
3.647
3.020
2.667
ꢀ2.496
ꢀ2.030
ꢀ0.933
2.322
1.988
0.847
—
5.438
—
—
1.767
—
—
1.611
—
16
Outliers (Z-scores above 2.5) are displayed in bold.
B. subtilis
S. aureus
a
b
8
6
6
R2 = 0.76
R2 = 0.72
4
2
4
2
0
0
-2
-1
0
1
2
3
4
5
6
-2
-1
0
1
2
3
4
5
6
7
8
-2
-2
exp. log (activity)
MRSA
exp. log (activity)
S. epidermidis
c
d
8
6
8
6
4
R2 = 0.78
R2 = 0.79
4
2
2
0
0
-2
-1
0
1
2
3
4
5
6
7
-2
-1
0
1
2
3
4
5
6
7
8
-2
-2
exp. log (activity)
exp. log (activity)
K. pneumoniae
e
8
6
4
2
0
R2= 0.77
-2
0
2
4
6
8
-2
exp. log (activity)
Figure 2. Performance of PLS models.
but less active (8-fold) against B. subtilis and inactive
against Gram-negative ones.
vsa_pol and Dipole_X (overall r2 ¼ 0:007, for com-
pounds 11a–p which are generally active r2 ¼ 0:17) sug-
gests a role for partial charges other than 5-nitro
imidiazole. Partial charge on carbonyl oxygen in benzo-
furan is about ꢀ0.6 in comparison with ꢀ0.16 on the
vsa_pol or polar fraction of van der Waals surface area
has a positive coefficient. Lack of correlation between

N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
Table 5. Details of best binary models
6361
Organism
Descriptor
Importance
A/I
a
a1
a0
S
S. aureus
Dipole_X
KierA2
ASA_P
0.50
0.28
0.28
12/19
94
83
100
3.4eꢀ5
5.4eꢀ6
3.4eꢀ5
5.4eꢀ6
5.3eꢀ6
S. epidermidis
MRSA
Dipole_X
Kier1
ASA_P
0.44
0.45
0.27
14/17
12/19
14/17
17/14
97
93
97
97
93
83
93
94
100
100
100
100
Dipole_X
KierA3
vsa_pol
0.50
0.13
0.28
B. subtilis
Dipole_X
KierA3
ASA_P
0.56
0.09
0.34
K. pneumonia
Dipole_X
Kier2
ASA_P
0.57
0.44
0.61
might suggest that there is limited space available for
this moiety of benzofuran system on target binding site.
Thus, analogues having small substituents, such as
chloro (11b), methyl (11e), and methoxy (11f), all exhib-
ited similar inhibitory activity relative to unsubstituted
benzofuran (11a), whereas the bulkier substituted ana-
logues bearing bromo (11c) and iodo (11d) substituents
proved to be inactive.
active inactive
Peformance of Binary Models
100
80
60
40
20
0
Among the different substituents on the C-5 position,
compound 11m having nitro substituent as an elec-
tron-withdrawing group revealed to be 2- to 8-fold less
active than unsubstituted compound (11a).
a
ilis
bt
oni
um
ne
.P
MRSA
Su
For the more potent compounds, the effect of changing
the position of substituent on the benzofuran moiety
was then investigated. Generally, changing the position
of substituents on the benzofuran appears to have little
influence (2- to 4-fold) on antibacterial activity, with the
exception for 6-chloro derivative (11g), shown to be
inactive against Staphylococci. Among these analogues,
compound 11n, having two strong electron-donating
group (OMe) at the 6 and 7 positions of the benzofuran
moiety, was the most potent against Gram-positive
and Gram-negative bacteria, with MIC values of 0.2–
0.78 lg/mL. This compound was less potent against
S. aureus and S. epidermidis, while exhibiting excellent
inhibition against MRSA, and B. subtilis superior than
reference drug, amoxicillin.
S. aureus
B.
S.epidermidis
K
Figure 3. Performance of binary models.
ether oxygen which is partially eclipsed in Z conformers
(especially in 7-substituted compounds as 11j–11l which
are generally active compounds), so it is likely that the
second important partial charge is the carbonyl group
of benzofuran system. Substitution of benzofuranone
moiety with indanone and indandione (15 and 16) re-
sulted in less to relatively inactive compounds. This phe-
nomenon shows that the oxygen of benzofuranone ring
could be the third important partial charge.
CASAꢀ (Negative charge weighted surface area, times
max negative charges)31 has negative coefficient, penaliz-
ing extra negative charge on phenol ring (for example
for nitro compounds 11m and 12m).
The most active compound, 11n, has low CASAꢀ, med-
ium Dipole_X and high value of Kier descriptors, sug-
gesting that large non-polar functional group on 7
position seems to be a good idea. KierA2, KierA3 are
also high for halogenated compounds 5-Br, 5-I and 5-
nitro compounds (generally inactive compounds) but
these compounds have higher CASAꢀ values.
Adding one of the alpha-shaped descriptors (KierA2,
KierA3) to previous two significantly increases the cor-
relation coefficients (r2 increase of 0:15–0:23) but when
used alone (data not shown), these predictors have very
low overall predictive power, suggesting their condi-
tional (rather than additive) role. KierA2 and KierA3
are highest for 5-Br, 5-I and 5-nitro compounds (gener-
ally inactive compounds) and lowest for 11a, 12a. This
Their spectrum of activity extended to penicillin-resis-
tant S. aureus makes these compounds attractive anti-
bacterial candidates also for treatment of infections
caused by organisms resistant to currently available

6362
N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
Table 6. Antimicrobial activity by 1–2, 11–12, and 15–16
Compound
R
MIC (lg/ml)
S. aureusa
S. epidermidisb
MRSAc
B. subtilisd
K. pneumoniaee
1
—
—
1.56
1.56
1.56
>100
6.25
12.5
>100
0.78
1.56
>100
>100
1.56
0.78
1.56
0.78
0.78
3.125
1.56
0.78
12.5
0.39
6.25
50
>100
>100
0.39
0.78
3.125
50
2
>100
3.125
3.125
>100
>100
3.125
3.125
>100
3.125
1.56
>100
3.125
3.125
>100
>100
3.125
3.125
>100
1.56
11a
11b
11c
11d
11e
11f
H
5-Cl
3.125
>100
>100
3.125
3.125
>100
1.56
5-Br
5-I
5-Me
5-OMe
6-Cl
0.39
0.39
0.78
0.39
0.39
1.56
0.39
0.78
6.25
0.2
11g
11h
11i
6-Me
6-OMe
7-Cl
1.56
1.56
11j
3.125
3.125
0.78
3.125
1.56
0.78
3.125
3.125
0.78
11k
11l
7-Me
7-OMe
5-NO2
6,7-(OMe)2
6-OHf
11m
11n
11o
11p
12a
12b
12c
12d
12e
12f
12.5
0.78
6.25
0.39
12.5
0.78
12.5
6.25
50
12.5
3.12
>100
>100
>100
>100
>100
>100
6.25
>100
>100
>100
>100
>100
>100
>100
0.2
f
4,6-(OH)2
H
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
50
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
50
5-Cl
5-Br
5-I
5-Me
5-OMe
6-Cl
6-Me
6-OMe
7-Cl
7-Me
7-OMe
5-NO2
—
12g
12h
12i
12j
12k
12l
12m
15
16
Amoxicillin
—
—
>100
>100
0.195
0.195
12.5
12.5
a Staphylococcus aureus ATCC 6538p.
b Staphylococcus epidermidis ATCC 12228.
c Methicillin-resistant Staphylococcus arueus (clinically isolated).
d Bacillus subtilis ATCC 6633.
e Klebsiella pneumoniae ATCC 10031.
f The antimicrobial assay was performed in the presence of BHT 0.1%.
drugs. Moreover, most of the 5-nitro analogues showed
excellent effectiveness against Gram-negative K. pneu-
moniae, better activity in respect to reference drug,
amoxicillin.
confirmed by leave one out cross validation method.
QSAR equations showed that Dipole_X, vsa_pol,
KierA2, and KierA3 correlate significantly with the anti-
bacterial activity, suggesting that possibly negative
charge of –NO2 group on C-5 of imidazole and partial
charge on carbonyl oxygen in benzofuran are necessary
for action. Moreover, there is a limited space available
for this moiety of benzofuran system, thus the bulkier
substituted analogues (11c–d) were inactive. Changing
the position of the substituent on the benzofuran ap-
pears to have little influence on the antibacterial activity.
In conclusion, a new series of substituted (Z)-2-arylid-
ene-3(2H)-benzofuranones possessing 1-methyl-5-nitro-
imidazole (11a–p) or 4-nitroimidazole (12a–m) were
synthesized and assayed for their antibacterial activity
against Gram-positive and Gram-negative bacteria.
Most of the 5-nitroimidazole analogues (11a–p) showed
a remarkable inhibition of a wide spectrum of Gram-po-
sitive bacteria (S. aureus, S. epidermidis, MRSA, and
B. subtilis) and Gram-negative K. pneumoniae, whereas
4-nitoimidazole analogues (12a–m) exhibited no effect
against selected bacteria. The quantitative structure–
activity relationship investigation was applied to find a
correlation between different physicochemical parame-
ters of the compounds studied and their biological activ-
ity. The predictive power of QSAR models built was
Acknowledgments
The authors sincerely thank Professor M. R. Ganjali for
his kind collaborations. This work was supported by
grants from the Research Council of Tehran University
of Medical Sciences and INSF (Iran National Science
Foundation).

N. Hadj-esfandiari et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6354–6363
6363
18. Spectral data for compounds 11a and 12a. (Z)-2-(1-
Methyl-5-nitroimidazole-2-ylmethylene)-3(2H)-benzofura-
none (11a). Yield, 56%; mp 239–241 °C; IR (KBr): 1716
References and notes
1. Public Health Laboratory Service. Commun. Dis. Rep.
2001, 11, 7.
(CO), 1481, 1383 (NO2) cmꢀ1 1H NMR (DMSO-d6) d
;
8.33 (s, 1H, H4-imid), 7.98–7.20 (m, 4H, Ar), 6.87 (s, 1H,
H-vinylic), 4.06 (s, 3H, NCH3); MS m=z (%) 271 (Mþ, 76),
225 (38), 184 (100), 155 (81), 127 (15), 101 (52), 76 (54), 50
(57). Anal. Calcd for C13H9N3O4: C, 57.57; H, 3.34; N,
15.49. Found: C, 57.40; H, 3.42; N, 15.63. (Z)-2-(1-
Methyl-4-nitroimidazole-5-ylmethylene)-3(2H)-benzofura-
none (12a). Yield, 60%; mp 261–264 °C; IR (KBr): 1716
2. Low, D. E.; Keller, N.; Barth, A.; Jones, R. N. Clin.
Infect. Dis. 2001, 32, S133.
3. Edwards, D. I. DNA-Binding and Nicking Agents. In
Enzymes and other Molecular Targets; Sammes, P. G.,
Taylor, J. B., Eds.; Comprehensive Medicinal Chemistry.
The Rational Design, Mechanistic Study & Therapeutic
Application of Chemical Compounds; Hansch, C., Ed.;
Pergamon Press: Oxford, 1990; Vol. 2, pp 725–751.
4. Edwards, D. I. J. Antimicrob. Chemother. 1993, 31, 9.
5. Knight, R. C.; Skolimowski, I. M.; Edwards, D. I.
Biochem. Pharmacol. 1978, 27, 2089.
6. Pires, J. R.; Saito, C.; Gomes, S. L.; Giesbrecht, A. M.;
Amaral, A. T. J. Med. Chem. 2001, 44, 3673.
7. Thomas, M. G.; Lawson, C.; Allanson, N. M.; Leslie, B.
W.; Bottomley, J. R.; McBride, A.; Olusanya, O. A.
Bioorg. Med. Chem. Lett. 2003, 13, 423.
(CO), 1501, 1347 (NO2) cmꢀ1 1H NMR (DMSO-d6) d
;
8.04 (s, 1H, H2-imid), 7.95–7.70 (m, 2H, Ar), 7.56–7.25 (m,
2H, Ar), 7.08 (s, 1H, H-vinylic), 3.82 (s, 3H, NCH3); MS
m=z (%) 271 (Mþ, 4), 225 (26), 157 (9), 151 (89), 120 (100),
105 (20), 92 (72), 68 (97). Anal. Calcd for C13H9N3O4: C,
57.57; H, 3.34; N, 15.49. Found: C, 57.71; H, 3.47; N,
15.30.
19. Varma, R. S.; Varma, M. Tetrahedron Lett. 1992, 33,
5937.
20. MOE. Chemical Computing Group Inc., Montreal, Que-
bec, Canada, 2006, see http://www.chemcomp.com.
21. Halgren, T. A. J. Comp. Chem. 1999, 20, 720.
22. Halgren, T. A. J. Comp. Chem. 1999, 20, 730.
23. Bashford, D.; Case, D. Annu. Rev. Phys. Chem. 2000, 51,
129.
24. Hall, L. H.; Kier, L. B. Rev. Comp. Chem. 1991, 2.
25. Ertl, P.; Rohde, B.; Selzer, P. J. Med. Chem. 2000, 43,
3714.
26. Wildman, S. A.; Crippen, G. M. J. Chem. Inf. Comput.
Sci. 1999, 5, 868.
27. Labute, P. Pac. Symp. Biocomput. 1999, 4, 444.
28. Baron, E. J.; Finegolg, S. M.; In Baily and Scott’s
Diagnostic Microbiology, 8th ed.; Mosby, C. V., St. Louis,
1990; pp 184–188.
8. Foroumadi, A.; Ghodsi, S.; Emami, S.; Najjari, S.;
Samadi, N.; Faramarzi, M. A.; Beikmohammadi, L.;
Shirazi, F. H.; Shafiee, A. Bioorg. Med. Chem. Lett. 2006,
16, 3499.
9. Foroumadi, A.; Oboudiat, M.; Emami, S.; Karimollah,
A.; Saghaee, L.; Moshafi, M. H.; Shafiee, A. Bioorg. Med.
Chem. 2006, 14, 3421.
10. Foroumadi, A.; Emami, S.; Hassanzadeh, A.; Rajaee, M.;
Sokhanvar, K.; Moshafi, M. H.; Shafiee, A. Bioorg. Med.
Chem. Lett. 2006, 15, 4488.
11. Foroumadi, A.; Emami, S.; Mehni, M.; Moshafi, M. H.;
Shafiee, A. Bioorg. Med. Chem. Lett. 2005, 15, 4536.
12. Foroumadi, A.; Pournourmohammadi, S.; Soltani, F.;
Asgharian-Rezaee, M.; Dabiri, S.; Kharazmi, A.; Shafiee,
A. Bioorg. Med. Chem. Lett. 2005, 15, 1983.
29. The antibacterial properties of 11o–p which were oxidiz-
able were performed in the presence of BHT 0.1% as an
antioxidant. The antibacterial activity of BHT 0.1% was
evaluated and no activity against any of the tested bacteria
was observed.
30. The Gram-negative bacteria utilized in this study consisted
of Escherichia coli ATCC 8739, Pseudomonas aeruginosa
ATCC 9027, and Klebsiella pneumoniae ATCC 10031.
31. Stanton, D.; Jurs, P. Anal. Chem. 1990, 62, 2323.
13. Tehranchian, S.; Akbarzadeh, T.; Fazeli, M. R.; Jamalifar,
H.; Shafiee, A. Bioorg. Med. Chem. Lett. 2005, 15, 1023.
14. Edwards, C. R.; Readhead, M. J.; Tweddle, N. J. J. Het.
Chem. 1987, 24, 495.
15. Robertson, A. J. Chem. Soc. 1932, 1389.
16. Carvalho, C. F.; Sargebt, M. V. J. Chem. Soc., Perkin
Trans. 1 1984, 1605.
17. Rufer, C.; Kessler, H.-J.; Schro¨der, E.; Damerius, A.
Chim. Ther. 1972, 7, 5.