Structure-dependent properties of wheat allelochemicals
Contemporary plant production systems rely heavily on inputs of chemicals to provide protection against diseases, insects and weeds. This extensive use of synthetic pesticides has given rise to concerns about chemical residues in the environment and importantly, the development of pesticide resistance in target organisms. Enhancing and harnessing defensive chemicals naturally produced by plants is an alternative approach which is gaining increased attention. The FATEALLCHEM project aimed to evaluate the possibilities of exploiting the allelopathic properties of wheat in conventional and organic farming systems through an assessment of target and non-target effects. For this purpose, researchers at the Istituto di Ricerche Farmacologiche Mario Negri developed a framework for the toxicity evaluation of allelochemicals based on models predicting their biological activity from structural information. QSARs (Quantitative Structure-Activity Relationships) were formulated based on the assumption that there is a quantitative relationship between the molecular structure of compounds and their biological, chemical and physical properties. The first step was to build up a data set of more than a hundred synthetic pesticides similar to the allelochemicals structure. Toxicity values were collected from different sources for different species, including daphnia, trout and ducks. Molecular descriptors that mathematically characterise their physicochemical behaviour were calculated using the Codessa software and Comparative Molecular Field Analysis (CoMFA). In the development phase of the modelling framework, different techniques were applied to extract the most relevant and often hidden information. Despite the lower performance, Group Method of Data Handling (GMDH) neural networks on the basis of chemical descriptors was found to be more general and applicable to heterogeneous structures. For CoMFA descriptors a different approach was used; Partial Least Square (PLS) indicated molecular metabolism processes of allelopathic toxicity. Furthermore, the comparison of these methods provided important information for better understanding the toxicity and action mechanism of allelochemicals. Researchers seek to collaborate with a suitable partner towards further validation of the QSAR models to ensure that they are adequate for making sufficiently accurate toxicity predictions for new compounds.
