Vol. 47 No. 2/2000 481–486 QUARTERLY

QSAR analysis of a set of previously synthesized 2,5,6-trisubstituted benzoxazole, benzimidazole and 2-substituted oxazolo(4,5-b)pyridine derivatives tested for growth inhibitory activity against Candida albicans, was performed by using the computer-assisted multiple regression procedure. The activity contributions for either heterocyclic ring systems or substituent effects of these compounds were determined from the correlation equation and the predictions for the lead optimization were described. The resulting QSAR revealed that the oxazolo(4,5-b)pyridine ring system with the substitution of a benzyl moiety at position 2 was the most favourable structure among the heterocyclic nuclei. Moreover, the fifth position in the fused ring system is found more significant than the other positions in improving the activity.

Most fungi are completely resistant to the action of antimicrobial drugs.Only a few substances have been discovered which exert an inhibitory effect on the fungi pathogenic for man, and most of these are relatively toxic (Meyers et al., 1976).
The synthesized compounds 1-61 used in this QSAR analysis indicated antifungal activity of MIC (minimum inhibitory concentration) values between 6.25-50 mg/ml against Candida albicans.Clotrimazole, oxiconazole and haloprogin were also used as standard drugs in order to compare the antifungal activity of these compounds.The observed data for the antifungal activity of the synthesized compounds and the standard drugs were found as given below (Ören, 1995): In this study, QSAR analysis of some novel antifungal benzoxazoles, benzimidazoles and oxazolo(4,5-b)pyridines 1-61 (Fig. 1) was performed by the MVA (multi variable analysis) technique.

EXPERIMENTAL SECTION
The MVA approach in QSAR analysis has been most widely and effectively used for the-oretical drug design due to various physicochemical (electronic, steric and hydrophobic) parameters and structural indicator parameters used together (Hansch & Fujita, 1964;1969).
The assumption can be formulated as given in Eqn.1: where I i is the structural indicator parameter and X j is the physicochemical variable.
In this study, the model is based on the in vitro activity of certain 2,5,6-trisubstituted benzoxazole, benzimidazole and 2-substituted oxazolo(4,5-b)pyridine derivatives 1-61 against C. albicans, where C is the MIC value expressed in molar concentration units (Table 1).
The variables used as descriptors in the analysis are electronic, steric and structural parameters.The structural indicator variable Ix expresses the replacement of -CH= by the isosteric group -N= in the six membered ring of the fused ring system.Ix is defined as 1 for -N= and 0 for -CH= in the compounds.The other indicator variable Iz has a value of 1 for the presence of a methylene group and 0 for its absence between the p-substituted phenyl moiety and the fused ring system in position 2. The indicator variable IR 2 has a value of 1 for a nitro group and 0 for a hydrogen atom at position 6 in the fused ring system (see Table 2).
Physicochemical parameters taken into consideration in QSAR study are F (field effect) as electronic influences and Verloop's STERIMOL parameters (L and B 4 ) for the steric interactions of the substituents R and R 1 (Hansch & Leo, 1979).L is defined as the length of a substituent along the axis of its substitution to the parent skeleton.B4 is defined as the maximum width of the substituent and might provide a better understanding of steric requirements in ligand-receptor interactions.Electronic effect of the  substituents, expressed in term of F, is found to be important in determining the activity, as it is predictive in electrophilic reactions of the drugs with the nucleophilic functions of biomolecules.
The correlation equation was performed using the Stepwise Technique by entering and removing the F level as 4.0 for each variable in the regression.
The overall F test value for the calculated best equation (Eqn.2) was found as 471.69 which is statistically significant at the level of probability P < 0.001 (Fig. 2), whereas the tabulated F value is 4.442.On the other hand, the predictive power of the performed QSAR model was also examined by calculating the q 2 (predicted r 2 ) value of some additional synthesized analog compounds (62-68) and the Eqn. 2 was found significant for the lead optimization in this set of compounds.

RESULTS AND DISCUSSION
The best equation observed in this QSAR study is: For the observation of the predictive power of the developed QSAR model, predicted log 1/C values of some additional analog compounds (62-68) were also calculated by using the correlation Eqn. 2 and compared with the observed values (Table 3).Calculated q 2 and S press /SSY values determine that the predictive power of this QSAR model is significant (Wold, 1991) as given below: q 2 = 0.67; S press = 0.65; S press /SSY = 0.22 < 0.4 Predictions for the lead optimization in this set of compounds can be concluded as follows.
Due to the activity contribution of every positive 0.3 value of regression coefficients of the structural indicator parameters Ix, Iz and IR 2 in the developed QSAR model, Eqn. 2 increases the antifungal activity as two fold improved potency.Holding a pyridine ring in the bicyclic system was found important for the heterocyclic fused system, providing two fold improved potency against C. albicans.Moreover, substituting position Z with a methylene group as a bridge element between the fused heterocyclic ring system and phenyl ring in this set of molecules also offered two-fold higher potency for the antifungal activity.However, having a nitro group at position R 2 in the bicyclic nucleus led to a nearly two-fold decrease in potency.
On the other hand, position R 1 was found more significant than the positions R and R 2 for the screened antifungal activity.On the basis of Eqn. 2, electronic and steric properties of the substituents on position R 1 were found to be more indicative than the hydrophobic effects.The developed QSAR model revealed that substituting position R 1 with a group possessing a positive field effect (F), concerning minimum length (L) and maximum width (B4), increases the antifungal activity.Consequently, the interpretation of numerical values of weights of these physicochemical descriptors obtained in Eqn.

Figure 2 .
Figure 2. Plot of the observed and calculated C values of the growth inhibitory activity of compounds 1-61 against C. albicans using Eqn. 2.

Table 1 . The structure and in vitro antifungal activity of the analyzed compounds 1-61 against C. albicans
a minimum inhibition concentration, b MIC value expressed in molar concentration 484 I • .Yalçìn and others 2000