ADDING EXTRA-DIMENSIONS TO PRIMARY METABOLOME PROFILING BY GC×GC-TANDEM IONIZATION TOF-MS INSIGHTS ON HAZELNUT (Corylus avellana L.) AROMA POTENTIAL

Introduction: This study focuses on hazelnuts (Corylus avellana L.) primary metabolome (i.e., amino acids, mono and disaccharides, low molecular weight acids and amines), and its characteristic distribution in fruits of different geographical origins and submitted to different post-processing practices. Its information potential, not only correlated with phenotype and pedo-climatic conditions, is of great interest to predict hazelnuts sensory quality after industrial roasting. Most of the (key)-aroma compounds and potent odorants [1,2] derives from non-volatile precursors during heating. Methods: Sample preparation consists in a selective extraction of primary metabolites (mostly polar compounds), followed by derivatization (oximation-sylilation) to obtain informative patterns by GC×GC with apolar × semi-polar column combination. The informative power is increased by using a time-of-flight mass spectrometry (TOF-MS) with tandem electron ionization as detector. Results: GC×GC patterns of derivatized primary metabolites account of about 530 peak-regions (most of the identified by matching linear retention index data with MS spectra signatures at 70 eV); the characteristic fingerprint enables sample clustering on the basis of key-variables: cultivars, geographical origin, drying and roasting condition, storage time. Samples with statistically relevant differences in the metabolite patterns, suggest a different aroma potential related to their geographical origin, predominant cultivar and post-harvest treatment. A dedicated methodology is proposed for sugars, primary metabolites playing a relevant role as aroma precursors. Longer 1D columns (60 m x 0.25 mm ID) enable better isomers separation and more confident identification; soft ionization energy (12 eV) produces spectra with a different information power for most of the informative analytes compared to those at 70 eV thanks to lower fragmentation rate and higher abundance of fragments with high m/z ratios. Conclusions: The analytical approach here proposed enables chemical fingerprinting of informative hazelnut primary metabolites, most of them showing coherent correlation with known thermal degradation (Maillard Reaction and Streker degradation - known pathways) volatile products. The approach can be adopted as a rational strategy for hazelnut quality assessment although its reliability and robustness have to be demonstrated with a meaningful sampling design covering the most influential variables. References 1. Hofmann, T., Schieberle, P. J. Agric. Food Chem. 46, 2721–2726 (1998) 2. Kiefl, J. & Schieberle P. J. Agric. Food Chem. 61, 5236–5244 (2013).