A. Tafi et al., MOLECULAR MODELING OF AZOLE ANTIFUNGAL AGENTS ACTIVE AGAINST CANDIDA-ALBICANS .1. A COMPARATIVE MOLECULAR-FIELD ANALYSIS STUDY, Journal of medicinal chemistry, 39(6), 1996, pp. 1227-1235
A series of 56 azole antifungal agents belonging to chemically diverse
families related to bifonazole, one of the antimycotic drugs of clini
cal use, were investigated using the comparative molecular field analy
sis (CoMFA) paradigm. The studied compounds, which have been already s
ynthesized and reported to be active in vitro against Candida albicans
, were divided into a training set and a test set. The training set co
nsisted of 40 molecules from all the different structural classes. Due
to the lack of experimental structural data on these derivatives, mol
ecular mechanics techniques were used to obtain putative active confor
mations for all the compounds. The correctness of this molecular model
ing work was confirmed a posteriori by comparison with structural data
of the analog 2w obtained by X-ray crystallographic analysis (Massa,
S.; et al. fur. J. Med. Chem. 1992, 27, 495-502). Two different alignm
ent rules of the training set molecules were used in this study and ar
e based on the assumption that according to published results on azole
antifungal agents, all the studied compounds exert their inhibitory a
ctivity through the coordination of their azole moiety to the protopor
phyrin iron atom of the fungal lanosterol 14 alpha-demethylase enzyme.
The predictive ability of each resultant CoMFA model was evaluated us
ing a test set consisting of 16 representative compounds that belong t
o all the different structural classes. The best 3D-quantitative struc
ture-activity relationship model found yields significant cross-valida
ted, conventional, and predictive r(2) values equal to 0.57, 0.95, and
0.69, respectively. The average absolute error of predictions of this
model is 0.30 log units, and the structural moieties of the studied a
ntifungal agents which are thought to contribute to the biological act
ivity were identified. The predictive capability of this model could b
e exploited in further synthetic studies on antifungal azoles. Further
more, the results obtained by using two different alignments of the in
hibitors suggest that the binding mode of these molecules involves bot
h a coordination to the iron protoporphyrin atom and an additional, li
kewise relevant, hydrophobic interaction with the active site.