This application note describes the use of Waters DART QDa System with Live ID to successfully differentiate between seemingly identical samples of turmeric spice based on differing levels of high value antioxidant anti-inflammatory chemoprotective active ingredients in the samples. The application will be of use to food safety scientists interested in food adulteration and authenticity studies, and quality control scientists testing cosmetic formulations.
Curcuma longa (C. longa) and Curcuma xanthorrhiza (C. xanthorrhiza) are plants belonging to the ginger family (Zingiberaceae), which are native to tropical South Asia but are widely cultivated in the tropical regions of the world. The dried rhizome of those plants, known as turmeric, is considered a traditional spice, dye and medicine, particularly in Ayurveda or traditional Chinese medicine.1 Currently turmeric is getting much attention since studies on its constituents have proposed many pharmacological properties (e.g. anti-oxidative, anti-inflammatory or anti-mutagenic properties,)2 although some of those constituents have recently been declared a PAIN (pan assay interference compound.)3 Extracts of C. longa and C. xanthorrhiza are also being increasingly used in cosmetic formulations as natural replacements for petrochemically derived ingredients offering antioxidant, anti-inflammatory, chemoprotective, and cosmetic coloring properties.4
The proposed pharmacological effects of turmeric are mainly attributed to so called curcuminoids, more precisely to the substances curcumin, demethoxycurcumin, and bisdemethoxycurcumin. While C. longa contains all three of those curcuminoids, C. xanthorrhiza lacks bisdemethoxycurcumin, and is therefore considered less valuable, which makes it cheaper on the international market. For this reason, commercially available powder of C. longa is often adulterated with C. xanthorrhiza, which causes financial damage for customers and food industry. This application note presents a fast and easy opportunity for the detection of such spice adulteration using DART-QDa analysis coupled with LiveID Software.
150 mg of dried and ground sample were weighed into 2 mL reaction tubes and mixed with 1 mL ACN:water 75:25 (v:v). After sonication for 15 min at 25 °C, the samples were centrifuged at 16800 rcf for 3 min. The supernatant was isolated and filtered through H-PTFE filters into 1.5 mL reaction tubes. The extracts were diluted 1:99 (v:v) with extraction solvent. Finally, each extract was spotted onto a QuickStrip card and analyzed using DART QDa.
The mass spectra of authentic samples of C. longa (n=10) and C. xanthorrhiza (n=10) were used to train a chemometric model. The parameters of the model can be seen in Table 1.
MS system: |
ACQUITY QDa |
MS source: |
DART SVP |
Ionization gas: |
He |
Ionization mode: |
DART +ve |
Gas temperature: |
450 °C |
Cone voltage: |
10 V |
Sample speed: |
1.00 mm/s |
Sampling frequency: |
10 Hz |
Acquisition mode: |
Full Scan |
Acquisition range: |
100–600 m/z |
Chromatography Software |
|
---|---|
MS software: |
MassLynx |
Informatics: |
Live ID sample recognition software |
The combined mass spectra were clustered in two groups using a Principle Component Analysis (PCA) combined with a Linear Discriminant Analysis (LDA). The resulting clusters can be seen in Figure 1.
Loading plots (Figures 2 and 3) show the significant ions contributing to the differentiation of classes. The ion at 369 m/z (curcumin) seems to be the major feature in PC1, accountable for approximately 70% of the variance. The ions at 217 (ar-turmerone) and 235 m/z (procurcumenol, isoprocurcumenol, curcumenol, curcumenone) are the main features in PC2 and responsible for approximately 16% of the variance. The corresponding structures are annotated in the spectra shown in Figure 4.
The created authenticity model was cross-validated using the “leave 1 file out” option provided by LiveID. The associated validation report can be seen in Figure 5. Out of 240 spectra, obtained from two different turmeric species, all could be associated with the correct species. This results in a correctness score of 100%. Additionally, the model was validated using the “leave 20% out” option provided by LiveID, which also resulted in a correctness score of 100%.
To further test the robustness of the model, two samples of turmeric powder were purchased from local supermarkets and extracted using the described method. The extracts were analyzed using the created LiveID authenticity model. Both samples could be recognized as C. longa and C. xanthorrhiza, respectively, with a correctness score of 100% (Figure 6).
The authors thank Arko Wicaksono for collecting and providing 20 authentic samples of turmeric used in this application note. This work was carried out in equal parts at the Institute of Analytical Chemistry and Radiochemistry of the University of Innsbruck and the Austrian Drug Screening Institute (ADSI) in Innsbruck, Austria under a collaboration agreement between the ADSI and Waters Corporation.
720007046, November 2020