Quantitative Structure-Activity Relationships of Sweet Isovanillyl Derivatives

Abstract

Isovanillyl derivatives constitute a large class of sweet compounds in which there is a high degree of structural similarity and a wide range of biological activity, the relative sweetness RS spanning from 50 to 10 000 times with respect to sucrose. This paper describes the results obtained by applying statistical models to develop QSARs for these derivatives. For a set of 14 compounds (set 1) appropriate physicochemical parameters for regression equations were selected using the genetic algorithm method. The best equation indicates a very close correlation (N 14, ND 5, r 2 0.982, r 2 cv 0:942, LOF 0.074, PRESS 0.271, S PRESS 0.184, S DEP 0.139). Good results have also been obtained by Molecular Field Analysis (MFA) applied to the same set of compounds (N 14, ND 4, r 2 0.957, r 2 cv 0:925, LOF 0.044, PRESS 0.348, S PRESS 0.196, S DEP 0.158). QSARs have also been derived for a larger set of 41 compounds (set 2, including set 1, plus other 27 compounds) with a much larger variety of structural types. These compounds have been divided into a training set of 35 compounds and a test set of 6 compounds. The most signi®cant QSAR obtained using physicochemical parameters (N 35, ND 6, r 2 0.673, r 2 cv 0:522, LOF 0.337, PRESS 7.432, S PRESS 0.515, S DEP 0.461) proved less successful than one using MFA parameters (N 35, ND 6, r 2 0.746, r 2 cv 0:607, LOF 0.261, PRESS 6.110, S PRESS 0.467, S DEP 0.418). PRESS values for the test set were 4.079 and 1.962 respectively showing that the MFA data had more predictive power.Equations with different numbers of descriptors were compared and it was concluded that the LOF which is dependent upon the number of parameters used as well as the sum of squares is a suitable measure of equation quality. These equations were also validated by scrambling the experimental data which gave signi®cantly worse agreement than the real data except when an excessive number of descriptors was used.

Computational studies of sweet-tasting molecules

“…An equivalent methodology was used for the sucrose and guanidine derivatives and the combined set of molecules, and this is equivalent to that detailed in our published work with the isovanillates [13]. For each study, the molecules were divided into a training set to develop the QSAR, and a test set to validate the QSAR.…”

“…The genetic algorithm was then used to select the most appropriate parameters to use in the equation using the methodology previously described [13]. Both the genetic factor algorithm (GFA) [18] and genetic partial least squares (GPLS) methods were used within the QSAR modules of Cerius 2 [14].…”

“…We derived QSARs for molecules in 4 families, the sucrose and guanidine derivatives discussed above together with sulfamate and isovanillate families. The molecules used in these latter 2 families, and their RS values are previously published [12,13]. There were 149 molecules in all, and these were divided into a training set of 120 molecules and a test set of 29.…”

“…QSARs can be established most readily with families of molecules, and we have studied several series with known sweetness value. In previous work, we have studied 21 sulfamates that have RS values of 0.6 to 70.5 [12] and 41 isovanillates with RS values between 1 to 10 000 [13]. For the sulfamates, we used principal component analysis to distinguish between molecules that tasted sweet and bitter, and also developed linear multiple regression equations with good predictive power.…”

“…For the sulfamates, we used principal component analysis to distinguish between molecules that tasted sweet and bitter, and also developed linear multiple regression equations with good predictive power. For the isovanillates, we also developed similar high-quality equations [13].…”