Electronic nose as rapid and innovative tool for the diagnosis of grapevine crown gall

ELECTRONIC NOSE AS A RAPID AND INNOVATIVE TOOL FOR THE DIAGNOSIS OF GRAPEVINE CROWN GALL The electronic nose (E-nose) is an innovative device developed for microbiological, food/drink and environmental safety, and medical applications. E-nose comprises appropriate sensor arrays, able to detect and discriminate several flavours and odours from different sources. In the last decade, E-nose is being applied to recognize plant diseases caused from different quarantine bacterial pathogens such as Erwinia amylovora, Ralstonia solanacearum and Clavibacter michiganensis subsp. sepedonicus. In grapevines, tumourigenic strains of Agrobacterium vitis cause crown gall disease which may give serious problems in nurseries and vineyards. For the first time, a portable E-nose (PEN3) was used to discriminate between galled and healthy vines (controls), experimentally inoculated with two different strains of A. vitis and water, respectively. E-nose data were related to the increase of inoculation site diameters measured 9 months after inoculation with the pathogen. Preliminary gas chromatography-mass spectrometry (GC-MS) analysis were performed to identify volatile species from grapevine samples and to optimize experimental conditions for sensorial analysis. Solid phase microextraction (SPME) CAR/PDMS fibers were used to concentrate trace level vapour emissions. Spectra from tumoured vines and controls had almost identical volatile compositions, except the key compound styrene detected in infected samples. In plants, styrene formation is compatible with decarboxylation of cinnamic acid involved in secondary metabolism of plants. All data (E-nose sensors responses and diameter increases) from samples were processed with Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). PCA application allowed a discrete separation of the samples into two groups, according to the presence or absence of disease symptoms. Moreover, the LDA model allowed the correct sample classification between healthy and galled vines (83.3%, cross validation). Tumours caused by the two strains of the pathogen were not discriminated according to the volatile compound composition identified by GC-MS. Although a larger number of grapevine samples should be analysed to create a more robust model, our results give interesting clues to go further with research on the diagnostic potential of this innovative system associated with multi-dimensional chemometric techniques. Keywords: E-nose; Agrobacterium vitis; GC-MS.