Abstract
This work assessed the antilisterial activity of a series of phenolic compounds, to improve understanding of the relationship between structure and biological activity allowing a more targeted approach to the use of plant antimicrobial ingredients in foods. Models were produced using classical QSAR that allows the assessment of the antilisterial activity of a congeneric group of phenols using multiple linear regression, neural network and fuzzy logic systems. The inhibition of a cocktail of 18 strains of Listeria monocytogenes by 25 mono-, di- and tri-substituted benzoic and cinnamic acids and 16 benzaldehydes was evaluated using the concentration, (C), required to give a 50% growth inhibition under anaerobic conditions at 35°C and pH 6.2 as a measure of biological activity (BAV). Using the method of least squares, multiple regression equations were obtained which described the contribution of some physicochemical and other structural properties of the compounds to their biological activity. The equation that best described the activity of benzoic and cinnamic acids was log(1/C)=3.54logK+0.93pKa-6.42, where K is a lipophilicity parameter and the effect of ionisation is represented by pKa. Benzaldehydes behaved differently, their activity being best described by the equation log (1/C)= -0.23Vw-0.44-1/E+0.25, where the activity is controlled by a steric parameter, the van der Waals volume (Vw), and an electronic-steric parameter for ortho substituents. Absence of a lipophilicity parameter indicates that partitioning into the cell membrane might not be required for antimicrobial activity. The models were tested in several food systems which showed that in food with a high protein or lipid content antilisterial activity was much lower than predicted, making the models unacceptable in such circumstances. The ability of artificial neural networks (ANN), fuzzy systems (FS) and multiple linear regression (MLR) to fit the biological activity surface describing the inhibition of Listeria monocytogenes by benzoic and cinnamic acid derivatives was compared. MLR and ANN were also compared for their ability to select the properties that best describe the biological activity of the compounds. The ANN method gave a better correlation, r2=0.96, compared with either MLR, r2=0.81 or FS, r2=0.92 and also a lower standard error, possibly indicating non-linearity in the data. The ANN was shown to generalise better than MLR using the leave-one-out method. The ANN selection algorithm for the selection of the parameters that contributed most to the biological activity of the phenols (logK and pKa.) agreed with the selected parameters of the MLR system. Aspects of the mechanism of action, resistance and effect of external factors on phenols activity were investigated using a Gram-negative organism (S. enteritidis), a Gram-positive acid resistant organism: (Lact. plantarum) and the organism of the study (L. monocytogenes). In general, Lact. plantarum showed a special resistance to the phenolic series tested. Its high ApH value, which is the proton motive force component most affected by the phenols, and its low content of surface sulfhydryl groups might explain these findings for the acids and the benzaldehydes respectively. S. enteritidis did not show marked resistance to the compounds, except for the cinnamic acid group. Since this groups main activity is explained almost entirely by their lipophilicity, the fact that Salmonella was the most lipophilic organism tested could explain this. The phenolic acids are compounds whose main activity is explained by their lipophilic character, but electronic aspects may be important, especially under aerobic conditions. This new behaviour was correlated in the case of L. monocytogenes to their antioxidant activity and only mono- or di-substituted benzoic acids with ortho OH substituents were found to be antioxidants. Therefore, it seemed reasonable to think that the electronic parameters found in the model describing activity under aerobic conditions account for free radical activity as well as for their dissociation state. No activity against hydroxyl radicals was detected for benzaldehydes or cinnamic acids. The benzaldehydes, on the other hand are compounds whose activity does not seem to depend on partitioning. Sulfhydryl groups on the outside of the cell are important targets for their action. All compounds tested were adversely affected by the presence of protein and lipids in the medium. Benzaldehydes were particularly susceptible in accordance with their proposed mechanism of action.