Geographical origin of guarana by LC-MS and chemometrics analysis

Paullinia cupana, variety sorbilis (Mart.) Ducke, popularly known as guarana, is a native Amazon plant species domesticated by Sataré-Mawé indians from the Maués region. It is considered to be one of most known Amazon forest products worldwide, because of its biological and medicinal properties. Despite its Amazon origin, today the biggest producer of guarana is the Bahia state, followed by Amazonas and Mato Grosso states. Guarana continues to be one of the most important non-timber products for the Amazonas state, due to the international market interest for the Amazon products. Therefore, it would be very important for the Amazonas economy that the guarana produced in the Amazonas state could have a certificated of origin based on a robust quality control method. Herein we report an analytical method based on metabolomics of guarana seed performed by LC-MS and chemometrics analysis.
Our research group was able to collect or acquire many authentic samples of guarana seeds from Amazonas, Bahia and Mato Grosso states. First, we developed a reproducible protocol that involved seeds pre-processing, sampling and fast microextraction process to ensure higher yields. Then we developed an LC-MS method, looking for the highest resolution and sensitivity, which was achieved using a Shimadzu UHPLC (Prominence model), coupled to a Bruker
Ion-Trap Mass Spectrometer (Amazon speed model). A PFP column from Phenomenex (Kinetex, 5 μm, 100 Å, 150x2 mm) was eluted with a gradient of aqueous formic acid (0.1%) and acetonitrile was selected. All samples from Amazonas (9), Bahia (8) and Mato Grosso (7) states were prepared at concentration of 1 mg.mL-1 and 5 µL were analyzed in triplicate. It was remarkable to observe how similar all chromatograms were, despite of their different origin, and solely chemometric analysis enable us to compare them. The obtained chromatograms were submitted to pre-treatment, first by peak alignment (COW algorithm), then smoothing, normalization and Pareto scaling. The data were analyzed by PCA and HCA analysis (Pirouette, Infometrix) for pattern recognition and they are presented below (Fig 1). We were able clearly to distinguish guarana from the three different local origins (blue for Amazonas, red for Bahia into two sets and orange for Mato Grosso). Further analysis let us identify the principal discriminants for each set and none of them were related to the major components of guarana. These results show the feasibility to have a robust quality control method that could certificate the origin of guarana from Amazonas.