Experimental and QSAR studies on antimicrobial activity of benzimidazole derivatives

Article abstract of DOI:10.1248/cpb.56.273

Twenty eight analogues of benzimidazoles had been synthesized and tested for their antimicrobial activity against four bacteria (Staphylococcus auerus, Escherichia coli, Bacillus pumilus and Proteus vulgaris) and two fungi (Aspergillus flavus and Aspergilus niger). Compounds with R as C ₆H₄NO₂ and R′ as SO₂C ₆H₄-CH₃(p), with R as C₆H ₄OCH₃ and R′ as SO₂C₆H ₄-CH₃(p), and with R as CH₂C₆H ₅ and R′ as CH₂(CH₂)₉Cl exhibited comparable or higher antibacterial activity than Ciprofloxacin against S. auerus and E. coli and, higher activity than Nystatin against A. flavus. Several other compounds showed better activity than the standard antibiotic for E. coli. Compounds with R as CCl₃ and R′ as SO ₂C₆H₄-CH₃(p) or COC ₆H₅ exhibited the lowest activity against all the organisms. Addition of methylene groups in the R′ position increased activity. Many of the compounds showed better activity than Ciprofloxacin for one or more organisms. Compound with R as CH₂OC₆H ₅ and R′ as CH₂(CH₂)₉Cl exhibited higher activity against both the fungii than the control Nystatin. Quantitative structure activity relationships (QSARs) developed were good for all the organisms (R²=0.65 to 0.88; R² _adj=0.63 to 0.86) and the predictive capability of the developed models was also reasonable (q²=0.52 to 0.83). The models had two to three independent variables. The data for the models which had three independent variables were divided into training and test/validation sets. The former set was used to develop the QSAR and these developed models were used to predict the activity of the test set data. In all the three cases the predictive capability of the models was good. The molecular descriptors identified were predominantly log P, electronic parameters, molecular size, shape and area. A positive correlation existed between the antibacterial activity and the first principal component.

Full text of DOI:10.1248/cpb.56.273

March 2008
Chem. Pharm. Bull. 56(3) 273—281 (2008)
273
Experimental and QSAR Studies on Antimicrobial Activity of
Benzimidazole Derivatives
a
b
,c
*
Ayarivan PURATCHIKODY, Govindasamy NAGALAKSHMI, and Mukesh DOBLE
a Department of Pharmaceutical Engineering and Technology, Anna University; Tiruchirappalli 620 024, India:
b Department of Pharmaceutical Chemistry, Erode College of Pharmacy; Erode 638 112, India: and c Department of
Biotechnology, Indian Institute of Technology Madras; Chennai 600036, India.
Received September 26, 2007; accepted December 13, 2007; published online December 25, 2007
Twenty eight analogues of benzimidazoles had been synthesized and tested for their antimicrobial activity
against four bacteria (Staphylococcus auerus, Escherichia coli, Bacillus pumilus and Proteus vulgaris) and two
fungi (Aspergillus flavus and Aspergilus niger). Compounds with R as C6H4NO2 and Rꢀ as SO2C6H4–CH3(p), with
R as C6H4OCH3 and Rꢀ as SO2C6H4–CH3(p), and with R as CH2C6H5 and Rꢀ as CH2(CH2)9Cl exhibited compa-
rable or higher antibacterial activity than Ciprofloxacin against S. auerus and E. coli and, higher activity than
Nystatin against A. flavus. Several other compounds showed better activity than the standard antibiotic for E.
coli. Compounds with R as CCl3 and Rꢀ as SO2C6H4–CH3(p) or COC6H5 exhibited the lowest activity against all
the organisms. Addition of methylene groups in the Rꢀ position increased activity. Many of the compounds
showed better activity than Ciprofloxacin for one or more organisms. Compound with R as CH2OC6H5 and Rꢀ as
CH2(CH2)9Cl exhibited higher activity against both the fungii than the control Nystatin. Quantitative structure
activity relationships (QSARs) developed were good for all the organisms (R2ꢁ0.65 to 0.88; R2adjꢁ0.63 to 0.86)
and the predictive capability of the developed models was also reasonable (q2ꢁ0.52 to 0.83). The models had two
to three independent variables. The data for the models which had three independent variables were divided into
training and test/validation sets. The former set was used to develop the QSAR and these developed models were
used to predict the activity of the test set data. In all the three cases the predictive capability of the models was
good. The molecular descriptors identified were predominantly log P, electronic parameters, molecular size,
shape and area. A positive correlation existed between the antibacterial activity and the first principal compo-
nent.
Key words benzimidazole; quantitative structure activity relationship; anti-infective activity; electronic descriptor; Log P; Ge-
netic function algorithm
With increase in the incidence of multidrug-resistant moniae18) and concluded that the more potent compound
gram-positive and gram-negative bacteria it becomes impera- would possess an oxazolopyridine ring system substituted
tive to continuously search for small molecules as anti-infec- with an electron withdrawing group at position 5 and benzyl
tive agents. Benzimidazoles fit this requirement well since moiety at position 2.19) They found that the antifungal activ-
they have demonstrated a diverse set of biological activities ity of these compounds against Candida albicans highly cor-
that include antibacterial, antiamoebic, antiviral, antifun- related with LUMO, molecular weight, R (resonance effect)
gal,1—4) anthelmintic,5) anti-HIV,6) antihistaminic,7—9) anti- and HOMO.20) Oxazolopyridine ring system with the substi-
ulcer,10,11) cardiotonic,12) antihypertensive13,14) and neurolep- tution of a benzyl moiety at position 2 exhibited activity
tic.15) They are also widely in clinical use. Their observed ac- against K. pneumoniae. Also a group which possessed hydro-
tivity depends upon the functional group attached to the moi- gen accepting capability improved the activity.21) In this
ety. In order to obtain more effective chemotherapeutic paper we report the synthesis of analogues of benzimidazole
agents, a variety of reports have been presented on the syn- with a variety of electron withdrawing groups at the Nꢀ1 po-
thesis and biological evaluation of new benzimidazoles.16)
A systematic structure based design may help in identify- two fungal species and the corresponding QSAR studies.
ing new benzimidazole analogues with higher activity. De- Modelling Methodology The structures of the twenty
sition, their antimicrobial activity against four bacterial and
veloping quantitative structure activity relationship (QSAR) eight benzimidazoles as listed in Table 1 was drawn and its
and using the relationship for designing newer structures energy was minimized by using Cerius2 software® (Acceryls
have shown reasonable success. QSAR studies on benzimi- Inc, U.S.A.) with Consistent Valence Force Field (CVFF)
dazole chloroaryloxyalkyl derivatives indicated that their in force field. This force field, also known as universal force
vitro activity against Salmonella typhi O-901 and Staphylo- field, is well suited for small organic molecules (not with
coccus aureus A 15091 depended on three descriptors compounds having metals) and it is parameterized using pep-
namely, HOMO energy, hydration energy and number of pri- tide and protein structures.22) Two hundred and forty nine
mary carbon atoms of the molecule.4) 2-Aminobenzimida- descriptors or structural features that include topological,
zole moiety connected to the 4-(5) position of an imidazole charge, geometrical, aromaticity, constitutive, quantum me-
ring through C2- or C3-methylene chains when tested on rat chanics and thermodynamics were evaluated for all these
brain membranes exhibited activity and it is found to have a structures. These descriptors describe the structural features
parabolic dependence on Log P.17) Activities of benzimida- of the molecule. Several literature reports give a very de-
zoles against Bacillus subtilis also indicated a parabolic rela- tailed description of these descriptors.23—25) Selecting a short
tion with Log P. Sener et al. have studied the QSAR of anti- set of descriptors to be used in the QSAR from this large
bacterial activity of benzimidazoles against Klebsiella pneu- pool is always a challenge. Several techniques have been re-
∗ To whom correspondence should be addressed. e-mail: mukeshd@iitmac.in
© 2008 Pharmaceutical Society of Japan

274
Vol. 56, No. 3
m

March 2008
275
ported in the literature for performing this task. A genetic al- nent analysis is performed on the descriptor data to reduce it
gorithm technique was used to select the descriptors from to a small set which can be analysed much more meaning-
this large pool that would give the best QSAR model. The fully.
genetic function algorithm performs a search over the space
Antimicrobial Activity The twenty eight compounds
of possible QSAR models using the lack of fit (LOFꢃdiffer- (see Table 1 for compound details) were tested against four
ence between the predicted and the actual activity values) bacteria namely, Staphylococcus aureus (ATTC-25923),
score to estimate the fitness of each model. Lack of fit indi- Bacillus pumilus (recultured), Escherichia coli (ATTC-
cates a difference between the model predictions and actual 25922) and Proteus vulgaris (recultured) and two fungi
values. Such evolution of a population of randomly con- namely, Aspergillus flavus (NCIM No. 524), and Aspergillus
structed models leads to the discovery of highly predictive niger (recultured). The activity was tested by the disc-diffu-
QSAR. The populations of the models are created by evolv- sion method under standard conditions using Müller–Hinton
ing random initial models using a genetic algorithm.26) This agar medium as described by NCCLS.30) Standardized inocu-
approach can build models using not only linear polynomials lum (5ꢄ105 cfu /ml) of each test bacterium was spread on to
but also higher-order polynomials, and other nonlinear func- sterile Müller–Hinton agar plates so as to achieve a confluent
tions. The QSAR is nothing but a multiple regression con- growth. Discs measuring 9 mm in diameter were punched
taining these selected descriptors as the independent vari- from Whatman no.1 filter paper. Batches of 100 discs were
ables and the activity as the dependent variable. The good- dispensed to each screw capped bottles and sterilized by dry
ness of the regression fits were estimated using parameters heat at 140 °C for 1 h. The test compounds (mol/ml) were
such as, R2 (ꢃ1ꢀSSE/TSS), R2
(ꢃ1ꢀ(nꢀ1)(1ꢀR2)/ dissolved in a mixture of EtOH/Me2SO4 mixture (1 : 1, v/v)
adj
(nꢀpꢀ1)), q2 (ꢃ1ꢀPRESS/TSS), and F ratio (ꢃ(nꢀ2)R2/ and the discs were introduced on the agar medium. Subse-
(1ꢀR2)) where, TSSꢃtotal sum of squares, PRESSꢃpredic- quently, the test compounds were loaded on the filter paper
tive sum of squares based on leave-one-out method.27) In the discs so that the disc contained 100 or 200 mg of the com-
leave-one-out method, regression relation is built by leaving pound. The discs only with EtOH/Me2SO4 mixture were used
one data point at a time. The model thus developed is used to as solvents control. The antibiotics namely, ciprofloxacin and
determine the activity of the data that was left out and the nystatin were used as standard drug to test the antibacterial
square of the difference between the experimental and model and antifungal activities respectively. The plates were al-
value is calculated. This exercise is repeated for all the data lowed to stand for 1 h or more for diffusion to take place and
points and the sum of square of the difference is estimated, then incubated at 37 °C for 24 h. The zone of inhibition was
which is known as PRESS. This method is also known as in- recorded to the nearest mm. Experiments were repeated in
ternal validation. While R2 and R2 are indication of the triplicate and the average values are reported here with their
adj
model fit, q2 is an indication of the predictive capability of standard deviation.
the model. A large F indicates that the model fit is not a
chance occurrence. R2 and R2adj above a value of 0.6 indicate Results
good model fit, while q2 above 0.50 indicates that the model
has reasonably good predictive capability.
The antimicrobial activity of the benzimidazole derivatives
is tested by the agar disc-diffusion method against the gram-
In order to further test the predictive capability of some of positive and gram-negative bacteria and, the two fungi. The
the models, the data set was divided into training and results of these studies are summarized in Table 2 with
test/validation sets. This division is made in a random man- the corresponding standard deviation from the mean value.
ner. The training set data was used for developing the regres- Two way ANOVA of the antimicrobial data at 100 and
sion model. This model was used to predict the activity of 200 mg/disc concentration levels for the four bacteria indicate
the test set data and compared with the actual values. This (Tables are given in supporting material) that the compounds
external validation technique definitely proves the predictive significantly differ in activity between themselves (Fꢃ10.2
capability of models and it is superior to internal leave-one- and 17.1 at 100 and 200 mg/disc concentration levels respec-
out method.
tively, pꢅ0.001*) and between the various microorganisms
Analysis of variance (ANOVA) on the activity data, cluster (Fꢃ2.83 and 5.83 for the two concentrations respectively,
analysis and principal component analysis were performed with corresponding probability values of pꢅ0.01** and
using KyPlot® (U.S.A.). Cluster analysis is an unsupervised pꢅ0.05*** respectively). Compounds 2b, with
R as
learning technique which aims at sorting different objects C6H4NO2 and Rꢁ as SO2C6H4–CH3(p), 2f with R as
into groups in a way that the degree of similarity (or activity) C6H4OCH3 and Rꢁ as SO2C6H4–CH3(p), and 4h, with R as
between two objects is maximal if they belong to the same CH2C6H5 and Rꢁ as CH2(CH2)9Cl exhibit comparable or
group and minimal otherwise.28) The forming of the groups higher antibacterial activity than ciprofloxacin against S.
is based on dissimilarity distant (i.e., differences between the auerus and E. coli and, higher activity than Nystatin against
activities of individual compounds). If two compounds are in A. flavus. The antibacterial activity of the compounds may be
the same group it means that they may have similar structural due to their direct binding to the double helix, thereby inter-
features. Compounds within a cluster are more similar to fering with the DNA-associated enzymatic processes.31)
each other than they are to compounds in other clusters. Prin- Compounds 2b and 2f have the same substitutent in the Rꢁ
cipal Components Analysis can be used to identify patterns position but have different groups in the R position. None of
in data, and expressing the data in such a way as to highlight the synthesised compounds exhibit higher antibacterial activ-
their similarities and differences. It is suited for high dimen- ity than control against B. pumilus and P. vulgaris. The clus-
sional systems, where the number of dimensions can be re- ter analysis of the activities of the compounds against S. au-
duced without much loss of information. Principal compo- reus at 100 mg/disc concentration is shown in Fig. 1. The

276
Vol. 56, No. 3
Table 2. QSAR Models for Various Bacterial and Fungal Activities (Number of Data Pointsꢃ28)
Concentration
Microorganism
Regression model
R2
R2
q2
F
PRESS
adj
(mg/ml)
S. aureus
E. coli
100
200
100
200
100
200
100
200
100
200
100
200
ꢀ2.793ꢆ0.0571 E-ADJ-mag ꢀ0.0361 Shadow-xy 3.971 A-type_O_58
ꢀ0.525ꢆ0.0427 E-ADJ-mag ꢆ0.00897 Shadow-xy ꢀ2.742 A type_O_58
ꢀ77.866ꢆ0.499 Molref ꢆ12.780 S_aaNꢀ0.0596 Jurs-WPSA-1
ꢀ70.64ꢆ0.525 Molref ꢆ10.759 S_aaN ꢀ0.0532 Jurs-WPSA-1
6.4069ꢆ2.292 Kappa-2 ꢀ0.1318 MR
0.75
0.68
0.719
0.64
0.854
0.866
0.75
0.626
0.745
0.64
0.58
0.63
0.77
0.72
0.62
0.52
0.83
0.83
0.72
0.58
0.68
0.56
0.53
0.58
23.98
17.2
53.8
59.26
43.8
23.6
27.29
16.7
19.489 95.399
36.1
48.1
0.871
0.881
0.78
0.653
0.77
0.676
0.609
0.66
0.794
0.744
24.72
32.68
65.94
146.9
15.5
B. pumilus
P. vulgaris
A. flavus
A. niger
5.926ꢆ2.286 Kappa-2 ꢀ0.101 MR
8.6ꢆ0.787 Sr ꢆ0.0688 Shadow-xy ꢆ0.26 LUMO
2.15ꢆ0.0965 Srꢆ0.18 Shadow-xyꢆ0.45 LUMO
12.489ꢆ1.042 A type_C_24 ꢀ61.295 Jurs-RNCG
12.5ꢆ1.16 A type_C_24 ꢀ61.02 Jurs-RNCG
3.305ꢆ1.831 A log p98ꢆ0.391 S-ss_O
2.789ꢆ2.12 A log p98ꢆ0.308 S-ss_O
110.8
24.53
84.15
40.38
67.4
0.741 47.6
0.67 36.4
Fig. 3. Effect of Adding Methlyene Group in the Rꢁ Position on the An-
tibacterial and Antifungal Activities (Compounds 4a to 4h, which Have
RꢃCH2C6H5)
Fig. 1. Cluster Analysis of Activity of S. aureus at a Concentration of
100 mg/ml
ꢀ, S. aureus; ꢁ, E. coli; ꢂ, B. pumilus; ꢄ, P. vulgaris; ꢃ, A. flavus; ꢄ, A. niger.
Clustering is based on group average.
of the antifungal azoles is attributed to the presence of ergos-
terol in the fungal cell membrane. The synthesized com-
pounds may block ergosterol synthesis by interfering with
the demethylation of its precursor, lanosterol.32) Compounds
1b with R as CCl3 and Rꢁ as SO2C6H4–CH3(p) and, 1a with
R as CCl3 and Rꢁ as COC6H5 generally exhibit the lowest ac-
tivity against all the organisms in this series. Cluster analyses
for other microorganisms are given in the supporting mate-
rial.
Figure 3 plots the effect of adding a methylene group
in the Rꢁ position (compounds 4a to 4h), which have
RꢃCH2C6H5 on the antibacterial and antifungal activities.
The antibacterial activity increases from 13—16 to 16—
20 mm and the antifungal activity from 9—10 to 17—18 mm
when the number of methylenes is increased from 3 to 10 in
the Rꢁ position of the benzimidazole derivatives, indicating
that increasing the hydrophobicity at the Rꢁ position favours
activity. For the addition of the same number of methylene
Fig. 2. Cluster Analysis of Activity of A. flavus at a Concentration of
100 mg/ml
Clustering is based on group average.
compounds in the higher activity group are marked with a groups in the Rꢁ position the increase in the antifungal activ-
circle. ity is higher than the increase in the antibacterial activity.
Compound 4a with R as CH2C6H5 and Rꢁ as CH2CH2- Figure 4 plots the effect of adding a methylene group in
CH2Cl exhibits lowest antifungal activity in this series at the Rꢁ position (compounds 3a to 3h), which have
both the concentration levels. Compound 3h, with R as RꢃCH2OC6H5 on the antibacterial and antifungal activities.
CH2OC6H5 and Rꢁ as CH2(CH2)9Cl exhibits higher activity The antibacterial activity increases from 12—15 to 16—
against both the fungii than the control Nystatin. Several 21 mm and the antifungal activity from 13—14 to 18—
compounds exhibit higher activity than the control against A. 20 mm when number of methylenes is increased from 3 to 10
flavus (see cluster analysis—Fig. 2). In general, the activity in the Rꢁ position of the benzimidazole derivatives. As ob-

March 2008
277
served before, for the addition of the same number of meth- pound has R equal to CCl3 and the most active compound
ylene groups in the Rꢁ position the increase in antifungal ac- has R equal to C6H4CH3. In the case of A. niger, the most ac-
tivity is higher than the increase in the antibacterial activity. tive compound also has C6H4CH3 group in the R position,
This observed increase in activity is higher for compounds while all other compounds have the same lower activity.
which have CH2OC6H5 in their R position, when compared
to those which have CH2C6H5 in their R position.
Table 2 lists the best QSAR models that were developed
using the 28 experimental data to describe the observed ac-
Figure 5 plots the effect of various electronegative func- tivity against the four bacteria and the two fungi at 100 and
tional groups in the position (ꢃCCl3, CH2C6H5, 200 mg/disc concentrations. Four of the linear regression re-
R
CH2OC6H5, C6H4NO2, C6H4CH3 and C6H4OCH3) when lations have two and the remaining eight of the equations
RꢁꢃSO2C6H4–CH3(p) on the antibacterial activities of the have three independent variables. The data fit for all the six
benzimidazole derivatives. The antibacterial activity of all microorganisms at both the concentrations are good with R2
the microorganisms correlate well with the highest occupied between 0.65 and 0.88 and R2 between 0.63 and 0.86. Ex-
adj
molecular orbital energy (with correlation coefficient values cept for two cases, the predictive capability of the models de-
of ꢀ0.62, ꢀ0.44, ꢀ0.67, ꢀ0.72 for S. aureus, E. coli, B. veloped for all the microorganisms are in the acceptable
pumilus and P. vulgaris respectively). More negative is the range (q2ꢇ0.56). The q2 for A. flavus at 100 mg/disc is 0.53
HOMO energy, lower is the activity. In the case of fungii, the and for S. aureus at 200 mg/disc is 0.52. The F values in all
activity does not correlate with HOMO but with a descriptor cases are very high and are statistically significant indicating
known as SIC (ꢀ0.79 and ꢀ0.73 for A. flavus and A. niger that the regression fit is not due to chance.
respectively), which describes the structural information. In-
In order to test whether the linear regression equations de-
creasing SIC (structural information content) decreases anti- veloped for S. aureus, E. coli and P. vulgaris at 100 mg/disc
fungal activity. In the case of A. flavus, the least active com- concentration were over fitted the data corresponding to
those organisms were divided into training and test/validation
sets containing 23 and 5 data points respectively. The train-
ing set data was used to develop the QSAR equations, which
are listed in Table 3. Later, these equations were used to pre-
dict the antibacterial activity of the test set. Figures 7 to 9
compare the predicted and experimental activities for the test
and training sets for all the three microorganisms. It is seen
in all the cases the test set data is predicted very well indicat-
ing the predictive capability of the models developed (the ta-
bles comparing the predicted with the actual values for these
three organisms are given in the supporting material).
Principal component analysis of the two hundred and forty
nine descriptors led to the conclusion that at least eight prin-
Fig. 4. Effect of Adding Methlyene Group in the Rꢁ Position on the Anti-
cipal components are required to describe 95% of the vari-
bacterial and Antifungal Activities (Compounds 3a to 3h, Which Have
ance, with the first principal component accounting for about
45% of the variance. Figure 10 plots the activities against the
RꢃCH2OC6H5)
ꢀ, S. aureus; ꢁ, E. coli; ꢂ, B. pumilus; ꢄ, P. vulgaris; ꢃ, A. flavus; ꢄ, A. niger.
Fig. 5. Effect of Adding Different Functional Groups in the R Position on Fig. 6. Effect of Adding Different Functional Groups in the R Position on
the Antibacterial Activity for Compounds with RꢁꢃSO2C6H4–CH3(p) as a the Antifungal Activity for Compounds with RꢁꢃSO2C6H4–CH3(p) as a
Function of HOMO
Function of SIC
ꢀ, S. aureus; ꢁ, E. coli; ꢂ, B. pumilus; ꢄ, P. vulgaris.
ꢀ, A. flavus; ꢁ, A. niger.
Table 3. QSAR Models for Three Bacterial Activities at a Concentration of 100 mg/disc Using Training Data set (Number of Data Pointsꢃ23)
Microorganism
Regression model
R2
R2
q2
F
PRESS
adj
S. aureus
E. coli
P. vulgaris
ꢀ3.735ꢆ0.0576 E-ADJ-mag ꢀ0.0278 Shadow-xy ꢀ3.992 A-type_O_58
ꢀ70ꢆ0.487 Molref ꢆ11.185 S_aaN ꢀ0.0563 Jurs-WPSA-1
8.72ꢆ0.732 Srꢆ0.0691 Shadow-xy ꢆ0.218 LUMO
0.734
0.872
0.732
0.693
0.853
0.69
0.64
0.82
0.67
1722
43.4
17.3
38.5
21.7
16.6

278
Vol. 56, No. 3
Fig. 7. Comparison of Activity Data and Model for S. aureus at a Concen-
tration 100 mg/ml
Fig. 9. Comparison of Activity Data and Model for P. vulgaris at a Con-
centration 100 mg/ml
ꢂ, training set data; ꢃ, test set data.
ꢂ, training set data, ꢃ, test set data.
Fig. 8. Comparison of Activity Data and Model for E. coli at a Concentra-
tion 100 mg/ml
Fig. 10. Antibacterial Activities Plotted against First Principal Component
ꢂ, training set data; ꢃ, test set data.
ꢀ, S. aureus; ꢁ, E. coli; ꢂ, B. pumilus; ꢄ, P. vulgaris.
Table 4. Details of Descriptors Used in the Models Listed in Table 2
Descriptor
Type
Thermodynamic
Atom type thermodynamic
Atom type thermodynamic
Information-content
Description
A log p98
Log P (the octanol/water partition coefficient) is related to the hydrophobic character of the molecule
Type of bonds attached to a carbon atom
Type of bonds attached to an oxygen atom (ꢃO)
A type-C-24
A type-O-58
E-ADJ-mag
Jurs -RNCG
Entropy of edge adjacency matrix
Shape and electronic information Charge of most negative atom divided by the total negative charge
Jurs-WPSA-1 Shape and electronic information Surface-weighted charged partial surface areas
Kappa-2
LUMO
Molref
MR
S_aaN
SC-1
Topological
Electronic
Thermodynamic
Fragment constants
Electrotopological
Topological
Kier’s Shape index
Lowest unoccupied molecular orbital
Molar refractivity
Molecular refractivity
N with 2 aromatic bonds
Number of the 1st order subgraphs
Shadow-xy
Sr
Shape
Electronic descriptor
Area of the molecular shadow in the XY plane
Captures changes in electronic distribution and it is an index indicating the reactivity of aromatic
hydrocarbons
S-ss_O
Electrotopological
O with two single bonds
four bacteria as a function of the first principal component. Log P (the logarithm of 1-octanol/water partition coefficient).
There seems to be a positive correlation between the antibac- The C log P is one such method which makes use of Ghose–
terial activity and the first principal component.
Crippen atom counts descriptors (A type).33,34) A type_O_58
The description of the various descriptors that were used is such a descriptor which represents a double bonded oxy-
in the QSAR is tabulated in Table 4. The descriptors used in gen in the form of ꢃO. Turker et al. observed that a group
the model for S. aureus are E-ADJ-mag, Shadow-xy and A- which possesses hydrogen accepting capability improves ac-
type_O_58. The first descriptor is positively correlated and tivity.21)
the other two are negatively correlated with activity. The first
The descriptors used in the QSAR for E. coli are Molref,
two descriptors relate to the molecular size and area. Hy- S_aaN and Jurs-WPSA-1. The first two descriptors have a
drophobicity of organic molecules is represented in terms of positive and the last one has a negative correlation with ac-

March 2008
279
tivity. Jurs descriptor is a combination of surface area and relating their observed activities with molecular descriptors.
surface charge of the molecule. Molref describes the molar All the models had between two to three independent vari-
refractivity and S_aaN captures the number of aromatic ni- ables. A genetic algorithm technique was used to select the
trogens in the compound. The descriptors used in the QSAR descriptors from a large pool so as to arrive at the best QSAR
for B. pumilus are Kappa-2 and MR. MR represents the mo- models. The QSARs developed were good for all the cases
lecular refractivity and is negatively correlated with the anti (R2ꢃ0.65 to 0.88; R2adjꢃ0.63 to 0.86). Except for two cases,
bacterial activity and Kappa-2 is positively correlated with the predictive capability of the developed models was in ac-
the activity. Kappa represents the shape of the molecule and ceptable range (q2ꢇ0.56). In order to further test the predic-
Kumar et al., similar to the current findings, also observed tive capabilities of the models that had three independent
positive contribution of Kier-shape index towards activity.35)
variables, the activity data for those cases were divided into
The descriptors used in the QSAR for P. vulgaris are Sr, training (23 values) and test/validation (5 values) sets.
Shadow-xy and LUMO, and all are positively correlated with QSARs were developed with the training data set and these
activity. Shadow-xy relates to the shape of the molecule and equations were used to predict the test data. In all the three
contribution of shape in QSAR for predicting activity against cases the models were able to predict the external test data
antimicrobials, specifically against Mycobacterium tubercu- well. The best molecular descriptors identified relate to polar
losis has been earlier reported by our research group.26,36) Sr surface area, log P, electronic parameters, molecular size,
and LUMO relate the electronic features of the compound. shape and area.
Sener et al. found that activity of similar compounds against
C. albicans is highly correlated with LUMO.20)
Experimental
The melting point of the synthesized compounds (Table 1) was deter-
mined in open capillaries using Toshniwal melting point apparatus. Purity
of the compounds was tested by TLC using an eluent mixture of
The descriptors used in the QSAR for A. flavus are A
type_C_24 and Jurs –RNCG. A type_C_24 represents the
carbon atom which is connected to two R groups on either
side though aromatic single bond.33,34) Such groups are pres-
ent in the present set of structures. Jurs –RNCG is negatively
correlated while the other descriptor is positively correlated
with activity. The Jurs descriptor is a combination of shape
and charge. The descriptors used in the QSAR for A. niger
are A log p98 and S-ss_O and both are positively correlated
with activity. The importance of Log P as a descriptor against
Bacillus subtilis has been pointed out by others.17) S-ss_O re-
lates to the electronic environment around the oxygen atom.
The models have identified several independent variables
or descriptors suitable for QSAR. They pertain to the size,
shape and area of the molecule, electronic parameters such as
Sr and LUMO, and liphophylic–hydrophilic balance. The de-
scriptors short listed in the present work match with those
that were found by several other researchers.14,15,18)
The present study included the synthesis of twenty eight
analogues of benzimidazoles which were tested for their an-
timicrobial activity against four bacteria namely S. aureus, E.
coli, B. pumilus and P. vulgaris and two fungi namely A.
flavus and A. niger. The test compound solutions were pre-
pared at two concentrations at constant molarity (mol/ml).
Detailed characterization data of sixteen of the compounds
are reported in this paper since they have not been mentioned
anywhere. Compounds 2b, with R as C6H4NO2 and Rꢁ as
SO2C6H4–CH3(p), 2f with R as C6H4OCH3 and Rꢁ as
SO2C6H4–CH3(p), and 4h, with R as CH2C6H5 and Rꢁ as
CH2(CH2)9Cl exhibited comparable or higher antibacterial
activity than Ciprofloxacin against S. auerus and E. coli and,
higher activity than Nystatin against A. flavus. Several other
compounds showed better activity than the standard antibi-
otic for E. coli. None of the synthesised compounds exhib-
ited higher antibacterial activity than control against B.
pumilus and P. vulgaris. A few compounds exhibited better
CHCl3 : CH3OH (3 : 1 v/v), and the structures were confirmed by FTIR (in
KBr pellet) using Perkin Elmer double beam Infra red spectrometer and 1H-,
13C-NMR (90 MHz) using EM-390-spectrometer. The mass spectra were
recorded on Shimadu QP-5000 mass spectrometer. Elemental analysis was
recorded on Perkin-Elmer 240 CHN analyzer. The reagents and solvents
were purchased from Ranbaxy, E-Merck, Loba and Sisco chemicals, India.
General Procedure A mixture of substituted benzimidazoles (1.0 g,
0.0042 mol) and appropriate dichloroalkanes (2.24 g, 0.15 mol) in aqueous
sodium hydroxide (5%, 20 ml) was stirred for 1—2 h at room temp. and
heated under reflux for 1—4 h. The reaction mixture was allowed to stand to
attain room temperature. A solid precipitate that separated out was filtered
off, washed with little cold water, dried in vacuum and recrystallized from
THF. Detailed characterization data is given below for compounds 3a—3h
and 4a—4h. The data for other compounds had been reported earlier by us
and are not reported here.29)
N-1-(Chloropropyl)-2-(phenoxymethyl)benzimidazole (3a): IR (KBr):
2968, 2818, 1635, 1602, 1476, 1446, 1206, 1069 and 734 cmꢀ1 1H-NMR
;
(DMSO-d6) d: 7.53—6.85 (m, 11H (Ar)), 4.78—4.34 (m, 6H (CH2)3) ppm;
13C-NMR d: 149.64, 144.02, 141.24, 133.64, 131.06, 129.83, 129.31,
128.83, 127.45, 120.61, 110.15, 108.83 and 100.31 ppm; MS (FAB) m/z:
300 (Mꢆ, 80%), 301 (Mꢆꢆ1, 40%), 299 (Mꢆꢀ1, 12%), 165 (16%), 115
(18%). Anal. Calcd for C17H17N2OCl: C, 68.00, H, 5.66, N, 9.33%. Found:
C, 68.06, H, 5.71, N, 9.37%.
N-1-(Chlorobutyl)-2-(phenoxymethyl)benzimidazole (3b): IR (KBr):
2907, 2717, 1642, 1612, 1486, 1426, 1208, 1066 and 736 cmꢀ1 1H-NMR
;
(DMSO-d6) d: 7.64—7.35 (m, 11H (Ar)), 4.85—4.79 (m, 8H (CH2)4) ppm;
13C-NMR d: 144.87, 142.22, 134.13, 129.78, 129.30, 128.87, 128.72,
128.37, 127.09, 124.24, 115.99, 113.29 and 102.23 ppm; MS (FAB) m/z:
314 (Mꢆ, 80%), 315 (Mꢆꢆ1, 40%), 313 (Mꢆꢀ1, 12%), 165 (16%), 115
(18%). Anal. Calcd for C18H19N2OCl: C, 68.78, H, 6.05, N, 8.91%. Found:
C, 68.83, H, 6.09, N, 8.97%.
N-1-(Chloropentyl)-2-(phenoxymethyl)benzimidazole (3c): IR (KBr):
2957, 2836, 1675, 1607, 1445 1408, 1208, 1069 and 740 cmꢀ1 1H-NMR
;
(DMSO-d6) d: 7.89—7.59 (m, 11H (Ar)), 4.85—4.79 (m, 10H (CH2)5) ppm;
13C-NMR d: 147.68, 146.89, 146.34, 135.70, 129.47, 128.46, 127.72,
121.58, 118.52, 115.58, 109.45, 104.40 and 101.58) ppm; MS (FAB) m/z:
328 (Mꢆ, 80%), 329 (Mꢆꢆ1, 40%), 327 (Mꢆꢀ1, 12%), 165 (16%), 115
(18%). Anal. Calcd for C19H21N2OCl: C, 69.51, H, 6.40, N, 8.53%. Found:
C, 69.55, H, 6.46, N, 8.57%.
N-1-(Chlorohexyl)-2-(phenoxymethyl)benzimidazole (3d): IR (KBr):
2982, 2811, 1660, 1593, 1449, 1400, 1212, 1605 and 736 cmꢀ1 1H-NMR
;
activity than the standard drug towards the fungus, A. flavus. (DMSO-d6) d: 7.93—7.62 (m, 11H (Ar)), 4.90—4.62 (m, 12H (CH2)6) ppm;
13C-NMR d: 143.23, 134.65, 130.94, 130.50, 129.58, 129.04, 128.37,
Compounds 1b with R as CCl3 and Rꢁ as SO2C6H4–CH3(p)
and, 1a with R as CCl3 and Rꢁ as COC6H5 generally exhib-
ited the lowest activity against all the organisms in this se-
ries. Addition of methylene groups in the Rꢁ position in-
128.08, 127.84, 127.54, 127.11, 110.94 and 104.65 ppm; MS (FAB) m/z:
342 (Mꢆ, 80%), 343 (Mꢆꢆ1, 40%), 340 (Mꢆꢀ1, 12%), 165 (16%), 115
(18%). Anal. Calcd for C20H23N2OCl: C, 70.17, H, 6.72, N, 8.18%. Found:
C, 70.22, H, 6.77, N, 8.23%.
N-1-(Chloroheptyl)-2-(phenoxymethyl)benzimidazole (3e): IR (KBr):
creased anti-infective activity. QSAR models were developed

280
Vol. 56, No. 3
2989, 2797, 1660, 1594, 1450, 1401, 1211, 1602 and 739 cmꢀ1 1H-NMR
;
N-1-(Chlorooctyl)-2-benzylbenzimidazole (4f): IR (KBr): 2968, 2889,
1642, 1611, 1444, 1403, and 736 cmꢀ1; 1H-NMR (DMSO-d6) d: 7.92—6.84
(m, 11H (2-benzylbenzimidazole)), 4.98—4.80 (m, 10H (CH2)8) ppm; 13C-
NMR d: 147.86, 140.94, 134.86, 132.23, 130.88, 129.93, 129.40, 129.23,
129.05, 128.70, 127.86 122.86 and 115 ppm; MS (FAB) m/z: 354 (Mꢆ,
100%), 355 (Mꢆꢆ1, 15%), 353 (Mꢆꢀ1, 10%), 296 (13%), 194 (4%), 178
(4%), 165 (18%), 115 (6%), and 105 (25%). Anal. Calcd for C22H27N2Cl: C,
74.57, H, 7.62, N, 7.90%, 2%). Found: C, 74.59, H, 7.67, N, 7.86%.
(DMSO-d6) d: 8.00—7.71 (m, 11H (Ar)), 4.96—4.62 (m, 14H (CH2)7) ppm;
13C-NMR d: 145.83, 134.14, 133.44, 130.20, 128.62, 127.22, 126.86,
124.84, 119.05, 116.82, 112.77, 109.05 and 104.72 ppm; MS (FAB) m/z:
356 (Mꢆ, 80%), 357 (Mꢆꢆ1,40%), 355 (Mꢆꢀ1, 12%), 165 (16%), 115
(18%). Anal. Calcd for C21H25N2OCl: C, 70.78, H, 7.02, N, 7.86%. Found:
C, 70.82, H, 7.07, N, 7.91%.
N-1-(Chlorooctyl)-2-(phenoxymethyl)benzimidazole (3f): IR (KBr):
2973, 2871, 1662, 1604, 1450, 1402, 1211, 1067 and 741 cmꢀ1 1H-NMR
;
N-1-(Chlorononyl)-2-benzylbenzimidazole (4g): IR (KBr): 2993, 2878,
1696, 1660, 1525, 1497, 1411 and 729 cmꢀ1 1H-NMR (DMSO-d6) d:
;
(DMSO-d6) d: 8.09—7.71 (m, 11H (Ar)), 4.99—4.63 (m, 16H (CH2)8) ppm;
13C-NMR d: 147.86, 140.94, 134.86, 133.05, 132.23, 130.88, 129.93,
129.40, 129.23, 129.05, 128.70, 127.86, 124.58 and 109.74 ppm; MS (FAB)
m/z: 370 (Mꢆ, 80%), 371 (Mꢆꢆ1, 40%), 369 (Mꢆꢀ1, 12%), 165 (16%),
115 (18%). Anal. Calcd for C22H27N2OCl: C, 71.35, H, 7.29, N, 7.56%.
Found: C, 71.39, H, 7.29, N, 7.55%.
7.00—8.00 (m, 11H (2-benzylbenzimidazole)), 4.98—4.82 (m, 10H (CH2)9)
ppm; 13C-NMR d: 149.20, 142.32, 133.07, 130.04, 129.22, 128.77, 127.58,
123.45, 116.80, 110.26, 105.58, and 104 ppm; MS (FAB) m/z: 368 (Mꢆ,
100%), 369 (Mꢆꢆ1, 20%), 367 (Mꢆꢀ1, 8%), 297 (29%), 178 (25), 165
(56%), 115 (38%), and 105 (60%). Anal. Calcd for C23H29N2Cl: C, 75.00, H,
7.88, N, 7.60%. Found: C, 75.03, H, 7.91, N, 7.65%.
N-1-(Chlorononyl)-2-(phenoxymethyl)benzimidazole (3g): IR (KBr):
2987, 2868, 1669, 1600, 1493 1449, 1214, 1069 and 740 cmꢀ1 1H-NMR
;
N-1-(Chlorodecyl)-2-benzylbenzimidazole (4h): IR (KBr): 3037, 2887,
1
1648, 1601, 1488 1461 and 738 cmꢀ1; H-NMR (DMSO-d6) d: 7.00—8.00
(DMSO-d6) d: 8.12—7.79 (m, 11H (Ar)), 5.09—4.69 (m, 18H (CH2)9) ppm;
13C-NMR d: 144.02, 141.24, 133.64, 131.06, 129.83, 129.31, 128.83,
127.45, 126.90, 121.78, 120.61, 112.15 and 110.44 ppm; MS (FAB) m/z:
384 (Mꢆ, 80%), 385 (Mꢆꢆ1, 40%), 383 (Mꢆꢀ1, 12%), 165 (16%), 115
(18%). Anal. Calcd for C23H29N2OCl: C, 71.87, H, 7.55, N, 7.30%. Found:
C, 71.88, H, 7.55, N, 7.30%.
(m, 11H (2-benzylbenzimidazole)), 5.01—4.82.(m, 10H (CH2)10) ppm; 13C-
NMR d: 148.92, 146.26, 133.41, 132.37, 132.02, 130.72, 129.41, 129.10,
128.92, 128.66, 128.34, 127.77, and 110.72 ppm; MS (FAB) m/z: 382 (Mꢆ,
100%), 383 (Mꢆꢆ1, 18%), 381 (Mꢆꢀ1, 9%), 194 (16%), 178 (15%), 165
(75%), 115 (15%), and 105 (38%). Anal. Calcd for C24H31N2Cl: C, 75.39, H,
8.11, N, 7.32%. Found: C, 75.43, H, 8.13, N, 7.34%.
N-1-(Chlorodecyl)-2-(phenoxymethyl)benzimidazole (3h): IR (KBr):
2975, 2888, 1642, 1612, 1444, 1215, 1407, 1071 and 742 cmꢀ1 1H-NMR
;
References
(DMSO-d6) d: 8.18—7.80 (m, 11H (Ar)), 5.15—4.72 (m, 20H (CH2)10)
ppm; 13C-NMR d: 144.47, 134.22, 134.13, 129.78, 129.30, 128.87, 128.72,
128.37, 127.09, 124.24, 115.99, 105.94 and 102.23 ppm; MS (FAB): m/z:
296 (Mꢆ, 100%), 297 (Mꢆꢆ1, 52%), 295 (Mꢆꢀ1, 13%), 193 (6%), 178
(13%), 165 (50%), 115 (10%) and 105 (13%). MS (FAB) m/z: 398 (Mꢆ,
80%), 399 (Mꢆꢆ1,40%), 397 (Mꢆꢀ1, 12%), 165 (16%), 115 (18%). Anal.
Calcd for C24H31N2OCl: C, 71.87, H, 7.78, N, 7.03%. Found: C, 71.91, H,
7.82, N, 7.03%.
1) Bishop B. C., Chelton E. T. J., Jones A. S., Biochem. Pharmacol., 13,
751—754 (1964).
2) Bhari N., Shailendra M., Garza G., Cruz-Vega D. E., Garza K., Saleem
F. Naqvi., Maura M. R., Azam A., Bioorg. Med. Chem. Lett., 12, 8—
11 (2002).
3) Devivar R. V., Kawashima E., Revankar G. R., Breitenbach J., Kreske
E., Drach J., Townsend L., J. Med. Chem., 37, 2942—2949 (1994).
4) Sheehan D. J., Hitchcock C. A., Sibley C. M., Clin. Microbiol. Rev.,
12, 40—79 (1999).
5) Saimo A. G., Cremieux A. C., Hay J. M., Meulemans A., Giovanangeli
M. D., Delaitre B., Coulaud F. P., Lancet, 17, 652—656 (1983).
6) Chimirri A., Grasso S., Monforte A. M., Monforte P., Zappala M., Il
Farmaco., 46, 925—933 (1991).
N-1-(Chloropropyl)-2-benzylbenzimidazole (4a): IR (KBr): 2967, 2852,
1
1660, 1594, 1479, 1449 and 735 cmꢀ1; H-NMR (DMSO-d6) d: 7.60—7.20
(m, 11H (2-benzylbenzimidazole)), 4.82—4.72 (m, 6H (CH2)3) ppm; 13C-
NMR d: 147.92, 146.26, 132.37, 132.02, 130.72, 129.56, 129.41, 129.10,
128.92, 128.66, 128.34, 127.77 and 126.48 ppm; MS (FAB) m/z: 284 (Mꢆ,
100%), 285 (Mꢆꢆ1, 17%), 283 (Mꢆꢀ1, 8%), 296 (4%), 194 (33%), 178
(21%), 165 (6%), 115 (31%) and 103 (25%). Anal. Calcd for C17H17N2Cl: C,
71.83, H, 5.98, N, 9.85%. Found: C, 71.84, H, 5.97, N, 9.89%.
7) Niemegeers C. J. E., Awouters F., Janssen P. A. J., Agents Actions, 18,
141—144 (1986).
8) Iemura R., Hori M., Ohtaka H., Chem. Pharm. Bull, 37, 962—966
(1989).
9) Benavides J., Schoemaker H., Dana C., Claustre Y., Delahaye M.,
Prouteau M., Manoury P., Allen J. V., Scatton B., Langer S. Z., Arbilla
S., Arzneim.-Forsch., 45, 551—558 (1995).
N-1-(Chlorobutyl)-2-benzylbenzimidazole (4b): IR (KBr): 2969, 2863,
1
1631, 1606, 1486 1400 and 737 cmꢀ1; H-NMR (DMSO-d6) d: 7.50—7.00
(m, 11H (2-benzylbenzimidazole)), 4.91—4.74 (m, 8H (CH2)4) ppm; 13C-
NMR d: 141.30, 139.83, 134.22, 134.13, 129.78, 129.30, 128.87, 128.37,
127.09, 124.24, 123.26 115.99, 113.29 and 105.94 ppm; MS (FAB) m/z: 298
(Mꢆ, 100%), 299 (Mꢆꢆ1, 19%), 297 (Mꢆꢀ1, 12%), 297 (2%), 194 (6%),
165 (13%), 115 (22%), and 103 (21%). Anal. Calcd for C18H19N2Cl: C,
72.48, H, 6.37, N, 9.39%. Found: C, 72.52, H, 6.41, N, 9.43%.
10) Ishihara K., Ichikawa T., Komuro Y., Ohara S., Hotta K., Arzneim.-
Forsch./Drug Res., 44, 827—830 ((1994).
11) Graham D. Y., Mccullough A., Sklar M., Sontag J. S., Roufail W. M.,
Stone R., Bishop R. H., Gitlin N., Cagliola A. J., Berman R. S.,
Humphries T., Digestive Diseases and Sciences, 35, 66—72 (1990).
12) Piazzesi G., Morano L., Ruegg J. C., Arzneim.-Forsch./Drug Res., 37,
1141—1143 (1987).
13) Wiedemann I., Peil H., Justus H., Adamus S., Brantl V., Lohmann H.,
Arzneim.-Forsch./Drug Res., 35, 964—969 (1985).
14) Kubo K., Kohara Y., Imamiya E., Sugiura Y., Inada Y., Furukawa Y.,
Nishikawa K., Naka T., J. Med. Chem., 36, 2182—2195 (1993).
15) Janssen P. A. J., Allewijn F. T. N., Arzneim.-Forsch./Drug Res., 18,
279—282 (1968).
N-1-(Chloropentyl)-2-benzylbenzimidazole (4c): IR (KBr): 2974, 2856,
1
1660, 1628, 1479, 1414 and 741 cmꢀ1; H-NMR (DMSO-d6) d: 7.64—6.91
(m, 11H (2-benzylbenzimidazole)), 4.93—4.71 (m, 10H (CH2)5) ppm; 13C-
NMR d: 149.45, 147.68, 146.89, 146.34, 135.70, 129.47, 128.46, 127.72,
127.23, 121.58, 118.52, 115.58, 109.45, 105.58, 100.46 and 21.53 ppm; MS
(FAB) m/z: 312 (Mꢆ, 100%), 313 (Mꢆꢆ1, 12%), 311 (Mꢆꢀ1, 7%), 297
(29%), 194 (25%), 178 (18%), 165 (88%), 115 (17%), and 105 (23%). Anal.
Calcd for C19H21N2Cl: C, 73.07, H, 6.73, N, 8.97%. Found: C, 73.11, H,
6.72, N, 8.97%.
N-1-(Chlorohexyl)-2-benzylbenzimidazole (4d): IR (KBr): 3065, 2858,
1
1638, 1602, 1479, 1414 and 739 cmꢀ1; H-NMR (DMSO-d6) d: 7.84—6.84
16) Nakano H., Inoue T., Kawasaki N., Miyataka H., Matsumoto H.,
Taguchi T., Inagaki N., Nagai H., Satoh T., Bioorg. Med. Chem., 8,
373—380 (2000).
(m, 11H (2-benzylbenzimidazole)), 4.95—4.74 (m, 10H (CH2)6) ppm; 13C-
NMR d: 145.62, 135.51, 131.32, 129.59, 129.50, 128.36, 127.32, 126.72,
125.20, 123.16, 114.10, 104.12 and 104.20 ppm; MS (FAB) m/z: 326 (Mꢆ,
100%), 327 (Mꢆꢆ1, 21%), 325 (Mꢆꢀ1, 10%), 295 (21%), 194 (8%), 178
(8%), 165 (8%), and 106 (16%). Anal. Calcd for C20H23N2Cl: C, 73.61, H,
7.05, N, 8.58%. Found: C, 73.65, H, 7.09, N, 8.62%.
17) Khalafi-Nezhad A., Soltani Rad M. N., Mohabatkar H., Asrari Z.,
Hemmateenejad B., Bioorg. Med. Chem., 13, 1931—1938 (2005).
18) Mor M., Bordi F., Silva C., Rivara S., Zuliani V., Vacondio F., Rivara
M., Barocelli E., Bertoni S., Ballabeni V., Magnanini F., Impicciatore
M., Plazzi P. V., Bioorg. Med. Chem., 12, 663—674 (2004).
19) Yalcin I., Sener E., Int. J. Pharm., 98, 1—8 (1993).
20) Sener E., Yalcin I., Sungur E., Quant Structure Activity Relat., 10,
223—228 (1991).
21) Turker L., Sener E., Yalcun I., Akbulut U., Kayalidere I., Sci. Pharm.,
58, 107—113 (1990).
22) Hobza P., Kabelac M., Sponer J., Mejzlik P., Vondrasek J., J. Comp.
Chem., 18, 1136—1150 (1997).
N-1-(Chloroheptyl)-2-benzylbenzimidazole (4e): IR (KBr): 2989, 2883,
1
1624, 1602, 1483, 1443 and 731 cmꢀ1; H-NMR (DMSO-d6) d: 7.89—6.86
(m, 11H (2-benzylbenzimidazole)), 4.96—4.78 (m, 10H (CH2)7) ppm; 13C-
NMR d: 145.64, 144.09, 129.43, 129.33, 128.91, 128.68, 127.31, 124.78,
115.44, 114.77, 110.31 and 105.42 ppm; MS (FAB) m/z: 340 (Mꢆ, 80%),
341 (Mꢆꢆ1, 16%), 339 (Mꢆꢀ1, 5%), 178 (4%), 165 (19%) and 104 (19%).
Anal. Calcd for C21H25N2Cl: C, 74.11, H, 7.35, N, 8.23%. Found: C, 74.13,
H, 7.32, N, 8.26%.

March 2008
281
23) Todeschini R., Lasagni M., Marengo E. J., Chemometrics, 8, 263—273
30) Yalcin I., Aki-Sener E., Int. J. Pharm., 98, 1—8 (1993).
31) Edwards D. I., J. Antimicrob. Chemother., 31, 9—20 (1993).
(1994).
24) Todeschini R., Consonni V., “Handbook of Molecular Descriptors,” 32) Borgers M., “The Scientific Basis of Antimicrobial Chemotherapy,”
Wiley-VCH, Weinheim, 2000.
25) Karelson M., “Molecular Descriptors in QSAR/QSPR,” Wiley Inter-
science, New York, 2000.
26) Sivakumar P. M., Geetha Babu S. K., Doble M., Chem. Pharm. Bull.,
55, 44—49 (2007).
27) Tarpey T., American Statistician, 54, 116—118 (2000).
ed. by Greenwood D., O’Grady F., Cambridge University Press, Cam-
bridge, 1985, pp. 133—153.
33) Viswanadhan V. N., Ghose A. K., Reyankar G. R., Robins R. K., J.
Chem. Inf. Comput. Sci., 29, 163—172 (1989).
34) Ghose A. K., Pritchett A., Crippen C. M., J. Comput. Chem., 9, 80—
90 (1988).
28) Everitt, Brian, Sabine Landau, Morven Leese, “Cluster Analysis,” 4th 35) Kumar A., Narasimhan B., Kumar D., Bioorg. Med. Chem., 15,
ed., Edward Arnold Publishers Ltd., London, 2001.
29) Puratchikody A., Nallu M., Gopalakrishna S., Indian J. Pharma. Sci.,
67, 47—52 (2005).
4113—4124 (2007).
36) Sivakumar P. M., Seenivasan S. P., Kumar V., Doble M., Bioorg. Med.
Chem. Lett., 17, 1695—1700 (2007).