Chemotaxonomy and phytochemistry of Heliantheae (Asteraceae) species and the use of Chemoinformatics in structure elucidation

Author(s)

Ricardo Stefani

• Document type: Master's Dissertation
• Press: Ribeirão Preto. , gráficos, ilustrações, tabelas.
• Institution: Universidade de São Paulo (USP). Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto (PCARP/BC)
• Defense date: 2002-10-02

Examining board members

Fernando Batista da Costa; Gil Valdo Jose da Silva; Paulo Cezar Vieira

• Advisor: Fernando Batista da Costa
• Field of knowledge: Physical Sciences and Mathematics - Chemistry
• Indexed in: Banco de Dados Bibliográficos da USP-DEDALUS
• Location: Universidade de São Paulo. Biblioteca Central do Campus de Ribeirão Preto; FFCLRP/Stefani, Ricardo

Abstract

Natural products chemistry has always been an important source for new andbioactive compounds. In modern world, mankind uses natural products to do many tasks: colouring, as essences, as agricultural defensives and many as medicines. Within the development of compound isolation techniques, the need for information organisation has grown. The need for quickly identification of isolated compounds has also grown. This was one of the necessities that made Chemoinformatics emerge. Chemoinformatics is a discipline that uses informatics as a tool to organise, analise and to generate new knowledge from chemical data. This tool has been used with success in automate structure elucidation, drug development (QSAR/QSPR) and to predict chemical-physical data of many molecules. The aims of the present work were the phytochemical study of species of the genera Dimerostemma and Ichthyothere to isolate new compounds, and the development of chemoinformatics techniques to aid natural products structure elucidation. The glandular trichome microsampling was made for diverse species of genera from the tribe Heliantheae (Viguiera, Tithonia, Dimerostemma). Many compounds were identified through glandular trichome microsampling. Two germacrolides and two eudesmanolides were identified from Dimerostemma species (D. brasilianum and D. episcopale), while from Ichthyothere terminalis two melampolides were identified, all of them being sesquiterpene lactones. Artificial Neural Networks were trained to make skeleton identification from data obtained from 13C NMR and the obtained results can be considered satisfactory. A software was developed to make automatic compound identification through the comparation with a compound library that possesses data from 51 STLs. This software is called NAPROSYS is also able to compare the NMR data of the sample with the NMR data stored into a compound library, making the imediate identification of compounds present into library possible and also help the structure elucidation of unknown compounds. To test NAPROSYS\' efficience to identify NMR data of compunds sored into the library was made with compounds isolated from species of Tithonia and Viguiera genera, because these genera has well describe compounds in the literature and that has been isolated in our laboratory, and the obtained results are excellent. Two Artificial Neural Network models were created to predict the retention time of sesquiterpene lactones in liquid cromatography (QSRR) with the aim of improve NAPROSYS performance in cromatographic data analysis. The results for this case, although coherent, can be improved. The conclusion of this work is that the use of classical techniques with the new techniques of chemoinformatics can be a very efficient tool to make structure elucidation, search for compounds with certain chemical properties and even the search for new bioactive compounds. (AU)