HPLC Analysis of Commercial Stevia Sweeteners

[seetharam annadana]

ELSD and UV - Complementary Detectors in the HPLC Analysis of Commercial Stevia Sweeteners

Introduction
Stevia, stevia rebaudiana bertoni, is a plant of the Compositae family native to Paraguay whose leaves have been used for centuries as a sweetener. Recently, it has been introduced as a crop in the United States and Canada in response to increased interest in natural foods. In its leaves, Stevia produces several sweet diterpene glycosides, which are non-glycemic, yet range in sweetness from 30 to 320 times that of sucrose. [1] Four major Stevia diterpene glycosides are recognized in the literature: stevioside, rebaudioside A, rebaudioside C and
dulcoside A. In addition to these, previously identified leaf constituents include volatile oil components, sterols, triterpenes, flavonoids, coumarins, and non-glycosidic diterpenes (sterebins). Research interest in Stevia has been oriented toward developing genetic lines in which levels of sweet-tasting glycosides are maximized and non-glycosidic
diterpenes are minimized. [1,2] Reversed phase HPLC with UV detection at 210nm has been used to determine constituents in commercial Stevia products, but the methodology is limited in its ability to separate all components of interest, especially when higher order oligosaccharides are present. Another problem is that any single UV wavelength is a
detection compromise as the constituents exhibit a wide range of absorbance maxima (193, 204, 236, 238, 284nm). Some are only weakly chromophoric. [3] The Evaporative Light Scattering Detector (ELSD) is a valuable complement to spectroscopic detectors for HPLC. The ELSD makes a measurement of photons scattered from semi- and non-volatile particles that have been dried of mobile phase through evaporation. Because its response is independent of the light absorbing
properties of molecules, it can reveal sample components that UV detectors miss and provide a more accurate profile of relative component abundance than is possible with a spectroscopic detector. [4]
For this study, a normal-phase methodology with complementary detectors, ELSD and multi-wavelength UV (PDA), reveals a wide range of constituents in commercial Stevia products in a single run.

Sample Preparation and Method
Commercial Stevia product A, a water solution, was prepared by using a 2:1 dilution with water. Commercial Stevia product B, a crystalline solid, was prepared by weighing 50mg into a 10mL volumetric flask
and diluting to the mark with water. The mixture was sonicated for 2 minutes. Stevioside was identified by authentic injection of the standard. Experiments using ELSD-LT evaporation temperatures from 28 - 40°C
showed no variation in detector response to the Stevia constituents, indicating that precise adherence to drift tube temperature is not critical to detection accuracy and reproducibility. Analytical Conditions
Mobile Phase: A: acetonitrile, B: 0.04% ammonium hydroxide
Gradient: (Time, B)(0,15)(60,65) Column: Shimadzu Premier Carbohydrate, 5mm, 250x4.6mm Injection Volume: 5mL
Flow Rate: 1mL/min. Column Temp: 30°C
Detector Settings Gain = 6; Temp = 40°C; Press. = 350kPa

Results and Conclusion
As shown in Figure 1 and consistent with the literature, UV detection at 210nm reveals the principal Stevia diterpene glycosides for commercial Stevia product B. The ELSD-LT, on the other hand, reveals a
complex mixture of chromophoric and non-chromophoric constituents in addition to the Stevia diterpene glycosides. The commercial additives erythritol and inulin fructo-oligosaccharides (FOS) are consistent with the
label claim. Erythritol and Dulcoside are co-eluting under these chromatographic conditions, the erythritol being detected by ELSD alone. When compared to low UV wavelengths, it is clear that the ELSD-LT delivers better baseline response for the gradient elution. Figure 2 shows comparison ELSD-LT profiles for the principal Stevia constituents of products A and B. The ELSD-LT reveals diterpenes, diterpene glycosides and non-terpenoid glycosides for Stevia product A in one run with a more accurate representation of constituent relative abundance. This is readily seen in Figure 3. Many of the minor constituents are weakly or non-chromophoric, making UV detection alone problematic when
the analyst is interested in more than the Steviol glycosides alone. Table 1 shows peak assignment and area % data for the Steviol glycosides of product A according to ELSD response. The % relative abundance of Steviol glycosides as determined by ELSD data is consistent with the literature. [1,2]

References
1. McGarvey, B. D., Attygalle, A. B., Starratt, A. N., Brandle, J. E., Xiang, B., Schroeder, F. C., Meinwald, J.,
Southern Crop Protection and Food Research Centre Poster of Research Efforts (1999).
2. Brandle, J. E., Starratt, A. N., Gijzen, M. Can. J. Plant Sci. 78, 527 – 536 (1998).
3. Semma, M., Fujii, M., Okada, Y., Kataoka, H., Ito, Y., Jpn. J. Food Chem., 8(2), 2001.
4. Young, C. S., Cereal Foods World., 47(1), 14-16 (2002).