Food Science & Safety

Dietary Supplements: Developing a ‘Fingerprinting’ Tool

Efforts are underway to establish a non-targeted quality control screen for plant extracts used as dietary supplements. Such a screen will improve the safety and efficacy to the consumer and meet the 2007 FDA ruling requiring cGMP compliance to ensure the quality of dietary supplements. Raw material from plant vary widely due to agricultural, harvesting, or processing methods. Intentional adulteration or non-intentional adulteration may also occur in the commercial product. To add further confusion, plant material used in dietary supplements is often identified generally by the common name (i.e. ginseng) and is not distinguished according to the species (i.e. Panax ginseng and Panax quinquefolius) or plant part (i.e. leaf, fruit, root, etc.) when used in the commercial product.

The large variation in potential components, composition and adulterants suggests that a non-targeted ‘fingerprinting’ method is needed to characterize a dietary supplement. NMR spectroscopy offers such capabilities and therefore is studied here as a potential screening tool.

NMR spectrum of various samples
NMR spectrum of various samples of Ginseng, Blueberry plant and Grape Seed extract showing the distinctive spectrum for each type of plant. The chemical shifts of the exchangeable protons showed high consistency between samples of a similar class and therefore provide a useful fingerprint of the phenolic region of the NMR spectrum. (b) 1D NMR spectra of Panax ginseng (North American Ginseng) and Panax quinguefolius.

Efforts NMR ‘Fingerprinting
The NMR spectrum of each sample type was distinctive and characteristic. Information available in the spectra were enhanced through use of a solvent (DMSO) that retains the exchangeable protons. Advantages to using NMR include:

  • Information rich (ability to view hundreds of molecules per spectrum)
  • Fully reproducible and fully quantitative

Ginseng Preliminary Screen
A Principal Component Analysis (PCA) of the NMR spectra from the extracts demonstrated that the NMR methods were suitable to distinguish various extract types. The view below of the first three PC’s shows clear separation of North American and Asian ginseng. Although results were only from a small sample set (5 samples) the details seen in the 1D spectra suggest that the samples from the two continents differ greatly and are readily distinguished . Additional samples of ginseng are currently been processed to more accurately define the sample variation for each region.

Grape Seed Vendor Analysis
The grape seed extract came from 18 different vendors. A PCA of the grape seed extracts by vendor showed that samples from 17 of the vendors grouped together whereas one vendor’s samples clearly differed from all other samples. Follow-up analysis revealed that the samples from vendor X were ‘beverage grade’ and were of lower purity than samples from other vendors.

Grape Seed Vendor Analysis
PCA (1,2,3) of crude natural product extracts

The PCA of Pine Bark Extract (PBE) and Grape Seed Extract (GSE) showed that one sample of PBE (PBE_3) was more similar to GSE than to PBE. Further analysis of the 1D spectrum of this sample suggested that this sample was in fact not GSE however it was also very unlike other PBE samples raising the question as to whether or not it had been adulterated or whether it was from a non-traditional pine bark extract source.

NMR spectra
(a) PCA (1,2,3) of PBE and GSE, (b) 1D NMR spectrum showed that PBE_3 had a very high concentration of catechin as is expected of GSE, however the flavonoid region (c) for the exchangeable protons showed that PBE_3 was not a GSE and was similar to PBE in flavonoid type and concentration.

Identification of Components
Individual components within the skullcap and germander samples were identified using a NMR Spectral Database (SBASE) match routine with AMIX Software (Bruker). This match routine used signal position and lineshape information to identify likely components in the crude mixture. Where possible, the presence of the suspected component was confirmed using a 2D MATCH with HSQC spectra. The fingerprint of the components from the 1D NMR in the crude spectra allowed prediction of the species of origin.

These preliminary data suggest NMR is a powerful tool to evaluate crude natural product extracts for information related to quality control, species identification, identification of components, and quantification of components. This nontargeted approach provides a quick assessment of samples that are not ‘typical’ and therefore suspect of poor quality, adulteration or mislabeling.