Optimization of an Untargeted DART-HRMS Method Envisaging Identification of Potential Markers for Saffron Authenticity Assessment

Abstract

Saffron is one of the most expensive agricultural products in the world and as such, the most commonly adulterated spice, with undeclared plant-based surrogates or synthetic components simulating color and morphology. Currently, saffron quality is certificated in the international trade market according to specific ISO guidelines, which test aroma, flavor, and color strength. However, it has been demonstrated that specific adulterants such as safflower, marigold, or turmeric up to 20% (w/w) cannot be detected under the prescribed approach; therefore, there is still a need for advanced and sensitive screening methods to cope with this open issue. The current investigation aims to develop a rapid and sensitive untargeted method based on an ambient mass spectrometry ionization source (DART) and an Orbitrap™high-resolution mass analyzer to discriminate pure and adulterated saffron samples with either safflower or turmeric. The metabolic profiles of pure and adulterated model samples prepared at different inclusion levels were acquired. Unsupervised multivariate analysis was carried out based on hierarchical cluster analysis and principal component analysis as first confirmation of the discriminating potential of the metabolic profile acquired under optimized DART-HRMS conditions. In addition, a preliminary selection of potential markers for saffron authenticity was accomplished, identifying compounds able to discriminate the type of adulteration down to a concentration level of 5%.

Keywords: DART; adulteration; authenticity; mass spectrometry; saffron; untargeted method.

Figure 1

Chronogram (a) along with mass spectra obtained by DART-HRMS analysis of pure saffron in negative (b) and positive (c) ion mode acquisition with the optimized conditions.

Figure 2

Hierarchical Cluster Analysis of the peak areas recorded for the compounds list obtained after filtration of the software output by p-value ≤0.05 in every ratio of the ANOVA test.

Figure 3

Hierarchical Cluster Analysis of the peak areas recorded for the compounds list obtained after filtration of the software output by p-value ≤0.01 in every ratio of the ANOVA test.

Figure 4

Principal component Analysis (PCA) scores plot (A) and loading plot (B) obtained for positive ionization mode after data filtration by a p-value threshold (≤0.05). Data points which strongly contribute to both principal components are target with arrows in the loading plot.

Figure 5

Signal intensity variation of candidate marker compounds among the inclusion levels of adulteration explored: 0 (pure saffron), 5, 10 and 20%: (A) safflower markers, (B) turmeric markers.