Identification of adulteration in botanical samples with untargeted mass spectrometry metabolomics

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Emily D. Wallace (Creator)
The University of North Carolina at Greensboro (UNCG )
Web Site:
Nadja Cech

Abstract: According to the 2017 Council for Responsible Nutrition survey, botanicals makeup 39% of the total dietary supplement usage in the United States. The use of dietary supplements in general has increased by 8% since 2015, and there is a need to ascertain and monitor the quality and authenticity of such products. Adulteration of dietary supplements is a concern because commercial suppliers may inadvertently or deliberately sell products for which composition does not match that reported on the label. Adulteration constitutes a potential health concern for consumers, increasing the risk of toxicity, adverse reactions, or ineffective products. Current methodologies employ targeted analysis or supervised statistical analysis for adulteration detection, both of which require prior knowledge of the sample set. A method for detection of adulteration in botanical dietary supplements utilizing untargeted mass spectrometry based metabolomics was developed and implemented on multiple instrument platforms. These included an ultraperformance liquid chromatography (UPLC) system coupled to a ultraviolet (UV) spectrophotometric detector (LC-UV), UPLC coupled to a quadrupole-time of flight (Q-ToF) mass spectrometer (LC-Q-ToF) and UPLC coupled to a hybrid quadrupole-orbitrap mass spectrometer (LC-Orbitrap). To evaluate the sensitivity of the method for detecting outliers, a set of samples was prepared by combining two different plant species, the botanical Hydrastis canadensis L. (Ranunculaceae), and a known adulterant species, Coptis chinensis Franch. (Ranunculaceae). C. chinensis was added to the H. canadensis samples in percentages ranging from 5% to 95% to emulate different levels of adulteration. The methodology was effective on all instrument platforms, but the sensitivity of detecting the adulterants varied depending on the analytical method and the method of data analysis. Using an unsupervised technique for data analysis (principal component analysis), the lowest percentage at which the adulterated sample was detectable as an outlier was measured based on the Hotelling’s T2 95% confidence interval. Outliers could be detected with this approach at 50%, 50%, and 10% adulteration using the LC-UV, LC-Q-ToF, and LC-Orbitrap systems, respectively. Composite score analysis was also performed for a statistical analysis comparison. A targeted analysis of a characteristic marker of adulteration (the alkaloid palmatine, which is a component of C. chinensis) was also conducted for comparison to the untargeted methods. Supervised statistical analyses, soft independent modelling by class analogy (SIMCA), was used to compare the sensitivity of different statistical approaches. The lowest percentage of adulteration detected as an outlier by these methods was 5%. SIMCA may be able to detect a lower percentage of adulteration, however, 5% was the lowest percentage tested in this study. The targeted analysis gave a limit of detection (LOD) of 0.0047 µM, 0.025 µM, and 0.027 µM; and a limit of quantitation (LOQ) of 0.12 µM and 0.55 µM, and 0.54 µM using liquid chromatography coupled to Orbitrap MS, Q-ToF MS, and Photodiode array (PDA) detectors, respectively. These values correspond to 0.3%, 1.5%, and 1.7% C. chinensis contamination in a botanical sample, respectively. Thus, a targeted methodology would detect trace levels of adulteration much more effectively than an untargeted method. However, untargeted methods have the added advantage of being applicable even when the identity of adulterants is unknown.

Additional Information

Language: English
Date: 2019
Adulteration, Goldenseal, Mass spectrometry, Natural products, Quantitation
Goldenseal $x Spectra
Dietary supplements $x Spectra
Natural products $x Spectra
Plant metabolites

Email this document to