Identification, Quantification, and Sensory Characterization of Steviol

Nov 13, 2014 - The chemical composition of four commercial S. rebaudiana extracts, obtained by different technologies, was characterized using UHPLC-E...
1 downloads 22 Views 864KB Size
Download PDF
Article pubs.acs.org/JAFC

Identification, Quantification, and Sensory Characterization of Steviol Glycosides from Differently Processed Stevia rebaudiana Commercial Extracts María Inés Espinoza,† Jean-Paul Vincken,§ Mark Sanders,§ Cristian Castro,† Markus Stieger,⊗ and Eduardo Agosin*,†,⊥ †

Centro de Aromas y Sabores, DICTUC S.A., Av. Vicuña Mackenna 4860, Macul, Santiago, Chile Laboratory of Food Chemistry, Wageningen University, P.O. Box 17, 6700 AA Wageningen, The Netherlands ⊗ Division of Human Nutrition, Wageningen University, P.O. Box 8129, 6700 EV Wageningen, The Netherlands ⊥ Department of Chemical and Bioprocess Engineering, School of Engineering, Pontificia Universidad Católica de Chile, Av. Vicuña Mackenna 4860, Macul, Santiago, Chile §

S Supporting Information *

ABSTRACT: Stevia rebaudiana is known for its sweet-tasting ent-kaurene diterpenoid glycosides. Several manufacturing strategies are currently employed to obtain Stevia sweeteners with the lowest possible off-flavors. The chemical composition of four commercial S. rebaudiana extracts, obtained by different technologies, was characterized using UHPLC-ESI-MSn. The composition of one of the ethanol-crystallized extracts (EC2) was entirely rebaudioside A, whereas the enzymatically modified (EM) extract contained the lowest concentration of this compound (2.7 mg/100 mg). The membrane-purified (MP) extract had the highest content of minor natural steviol glycosides (23.7 mg/100 mg total extract) versus an average of 2.4 mg/100 mg total extract for the EC samples. Thirteen trained panelists evaluated sweetness, bitterness, licorice, and metallic attributes of all four extracts. The highest licorice intensity (p ≤ 0.05) was found for MP. Both samples EC1 and EC2, despite their different chemical compositions, showed no significant differences in sensory perception. KEYWORDS: Stevia extracts, UHPLC ESI MS, natural sweeteners, sensory evaluation

■

INTRODUCTION

For decades, synthetic non-nutritive, high-intensity sweeteners such as saccharin and aspartame have been used to replace sugar and other caloric sweeteners. In recent years, much effort has been directed toward the development of noncaloric sweeteners from natural sources, to meet increasing consumer demand for organic products. Stevia rebaudiana Bertoni is a natural source for non-nutritive sweeteners, native to Paraguay and Brazil.1 The Stevia plant has been used for centuries in South America and for decades in Japan and other Asian countries. Since 2008, rebaudioside A, the major component of Stevia leaf extract, has been recognized as safe (GRAS) by the U.S. Food and Drug Administration (FDA) and permitted as food additive;2 since December 2011, the sale and use of steviol glycosides as food additive is also permitted in the European Union.3 Stevia consists of a number of diterpenic ent-kaurene glycosides that have an intense sweet flavor, but the magnitude and quality of the taste differ between molecules.4 Most of the diterpenes have the steviol aglycone (13-hydroxy-ent-kaur-16en-19-oic acid; skeleton 1; Figure 1) with different sugar moieties.5,6 The quantitatively dominant steviol glycosides are stevioside (Sv) and rebaudioside A (Ra). Recently, the occurrence of additional diterpene glycosides with a slightly modified kaurene skeleton (skeleton 2; Figure 1) was reported.6,7 © XXXX American Chemical Society

Figure 1. Structures of steviol glycosides. R1 and R2 represent different sugar moieties. R3 = CH2OH.

The composition and quality of purified Stevia extracts ultimately depend on the extraction process employed by the manufacturer. Some commercial extracts reflect the natural composition of steviol glycosides in the plant, whereas others do not. Differences in composition between extracts might become an issue with respect to regulation, sensory perception, and consumer preferences. Different strategies to improve the sweetening properties of the final extract have been developed.8 Some manufacturing strategies are targeted toward obtaining the glycosidic fraction by using organic solvents, whereas other Received: June 20, 2014 Revised: November 13, 2014 Accepted: November 13, 2014

A

dx.doi.org/10.1021/jf502878k | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

parent ion and MSn spectral data. The most abundant ion of MS2 was used for MS3. To visualize the minor compounds, the analyses were carried out with a divert valve (DV), by which the main peaks (Ra and Sv) could be removed from the chromatograms. The injection volumes were 1 and 5 μL for analysis without or with DV, respectively. Quantification of Ru, Sb, Rb, Rc, Rd, Rf, and DuA was conducted in SRM mode. Due to their coelution and overlapping fragments in MS2, Ra and Sv were quantified in single ion monitoring (SIR) mode, as the rest of the minor compounds, using the corresponding [M − H]− ion. Calibration Curves. The standard compounds were accurately weighed and dissolved in MeOH/H2O (50:50 v/v) to obtain stock solutions with a final concentration of ∼0.1 mg/mL. Dilutions were prepared in MeOH/H2O (50:50 v/v) in the range of 0.005−1 μg/mL for Ru, Sb, Rb, Rc, Rd, Rf, and DuA and in the range of 0.1−10 μg/mL for Sv and Ra. Three independent determinations were conducted for each concentration. Calibration curves were calculated by linear regression of double-logarithmic plots of the area of the selected ions versus the concentration of the respective steviol glycoside standard. These calibration curves were employed to calculate the amount of major and minor steviol glycosides present in the original samples. The obtained linearity coefficients were >0.998 (see Table S1 in the Supporting Information). For some steviol glycosides, no commercial standards were available. Thus, Rd was used for the quantification of minor steviol glycosides and transglycosylated products containing the Glcβ1−2(Glcβ1−3)Glcβ1- chain; and Rc was used for the quantification of compounds containing the Rhaα1−2(Glcβ1− 3)Glcβ1- chain. Sensory Analysis. The sensory attributes ‘sweetness’, ‘bitterness’, ‘licorice’, and ‘metallic’ of the four commercial samples of Stevia (MP, EM, EC1 and EC2) were assessed by a trained panel that was previously recruited by the Centro de Aromas y Sabores following ISO norms.20 All panelists (n = 13) had previous experience in sensory analysis of other products. The panel received a specific training to assess the sensory properties of the Stevia extracts and employed a category-specific scale of 100 mm (product specific) for the sensory assessment.21 Sensory data were acquired with the software Compusense Five release 5.4 (Guelph, Canada) and analyzed (ANOVA, means, SD) with Senstools (version 3.0.11, OP&P Product Research, Utrecht, The Netherlands). Selection and Training of the Panelists. The selection procedure consisted of two phases. First, 33 female candidates 45 ± 5 years of age were asked to rank seven randomly presented sucrose solutions (2, 3, 5, 7, 9, 11, and 13% w/v sucrose) in order of increasing sweetness. In the second phase, the panelists had to position samples of sorbitol (86200 mg/L), sucralose (78.6 mg/L), and commercial Stevia (276 mg/L) within the correctly ordered sucrose solutions. These samples were designed to match the sweetness of a 5% w/v sucrose solution, as proposed elsewhere.22,23 All of the sessions were performed in standardized sensory booths. Among the candidates, 13 panelists that performed correctly both tests were selected for further training. The selected panelists completed five introductory training sessions of 2 h each during a 2 month period. First, they were introduced to the main sensory attributes describing sweeteners (sweetness, bitterness, licorice, and metallic taste). For this purpose, five solutions were shown to the panelists to familiarize them with the attributes: sucrose (sweetness), commercial Stevia (sweetness), caffeine (bitterness), glycyrrhizic acid ammonium salt (licorice), and iron sulfate (metallic taste). In a second phase, a product-specific scale was developed to reflect the typical bitterness, licorice, metallic, and sweet tastes of the solutions that were assessed later by the panelists. The intensity range of each attribute was assessed on a scale from 0 to 100 mm. For each attribute, a reference stimulus was provided, which was anchored with an intensity in the middle of the scale at 50 mm. A series of 5% w/v sucrose solutions with increasing concentrations of tastants were presented to estimate the concentration of the reference stimulus, such as its perceived intensity, was similar to the intensity of each attribute present in the four Stevia samples. The reference stimuli for each

strategies prefer an aqueous purification process, where neither the glycosides are modified nor organic solvents are included in the purification.9 Alternatively, some strategies focus on the isolation of specific constituents, such as rebaudioside A (Ra).10 Moreover, enzymatic transglycosylation of stevioside and its congeners has been used by some manufacturers to improve the sweetness and reduce the undesirable bitter aftertaste of these compounds. Often, cyclodextrin glycosyltransferases (CGTases) have been used for this purpose. CGTases transfer glycosyl residues from donor substrates, such as cyclodextrins11 or starch, to the C13 and/or C19 sugar residues of the natural compounds, yielding a complex mixture of transglycosylated products.12,13 Ultrahigh-performance liquid chromatography (UHPLC) is the method of choice for compositional analysis of steviol glycosides in extracts. Different kinds of detectors have been used for identification and quantification: UV,14,15 pulsed amperometric detection16 (PAD), charged aerosol detector (CAD),17 and mass spectrometry (MS).18,19 The aim of this study was to determine the chemical composition of the major and minor steviol glycosides by ultrahigh performance liquid chromatography−electrospray ionization−mass spectrometry (UHPLC-ESI-MS) of four commercial Stevia products. In addition, the sensory properties of the latter were evaluated by a trained panel to assess the impact of the different manufacturing processes and resulting chemical composition on the sensory quality of the extracts.

■

MATERIALS AND METHODS

Chemicals and Materials. Stevioside (Sv), rubososide (Ru), steviolbioside (Sb), rebaudioside A (Ra), rebaudioside B (Rb), rebaudioside C (Rc), rebaudioside D (Rd), rebaudioside F (Rf), and dulcoside A (DuA) were purchased from ChromaDex (Irvine, CA, USA); their purity was >97%. UHPLC-MS grade water and acetonitrile (ACN) were purchased from Biosolve BV (Valkenswaard, The Netherlands). The following materials were used: sucrose (Merck, Darmstadt, Germany), commercial Stevia (Daily, Santiago, Chile), caffeine (Sigma-Aldrich, Germany), iron sulfate (Merck), D-sorbitol 98% (Sigma-Aldrich), sucralose (Sugafor, Santiago, Chile), glycyrrhizic acid ammonium salt from Glycyrrhiza uralensis root (Sigma-Aldrich), and unsalted rice cookies (Jumbo, Santiago, Chile). Stevia Samples. Purified S. rebaudiana extracts obtained by different process technologies, that is, membrane purification (MP), enzymatic modification (EM), and two samples from ethanol crystallization (EC1 and EC2), were provided by Prodalysa (Concón, Chile), AccoBio (Jiangsu, China), GLG Life Tech Corp. (Vancouver, Canada), and Cargill Truvia (Minneapolis, MN, USA), respectively. UHPLC-DAD-ESI-MSn Analysis. Samples were analyzed in a Thermo Accela UHPLC system (Thermo Scientific, San Jose, CA, USA) equipped with a pump and autosampler using an Acquity UPLC HSS C18 column (Waters; 150 mm × 2.1 mm i.d.; particle size = 1.8 μm) at 20 °C. The eluents used were water/acetic acid (100:0.1, v/v) (eluent A) and ACN/acetic acid/dichloromethane (100:0.1:0.1, v/v/ v) (eluent B). The elution program was as follows: 0−1 min, 25% B (v/v); 1−16 min, linear gradient of 25−50% B (v/v); 16−22 min, linear gradient of 50−100% B (v/v); 22−28 min, linear gradient of 100−25% B (v/v). The flow rate was 300 μL/min. MSn analysis was performed on a Thermo Scientific LTQ VELOS, using electrospray ionization (ESI) and detection in negative mode, with a source voltage of 3.5 kV and an ion transfer tube temperature of 350 °C. The instrument was tuned manually to optimize the ionization process and sensitivity using rebaudioside A. Control of the instrument and data processing were carried out using Xcalibur 2.07 (Thermo Scientific). Identification of compounds was based on the molecular mass of the B

dx.doi.org/10.1021/jf502878k | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

Statistical Data Analysis. The means and standard deviations were calculated using the software Senstools. A repeated-measures ANOVA was used with samples as factor and judges to test for the significance of differences. A LSD post hoc comparison was performed. A significance level of p < 0.05 was chosen.

sensory attribute, with their respective concentrations, were as follows: caffeine for bitterness (0, 65, 110, 155, 210, and 250 mg/L), glycyrrhizic acid ammonium salt for licorice (100, 300, 500, 700, 900, and 1000 mg/L), and iron sulfate for metallic (0, 8, 10, 12, and 14 mg/ L). The panelists had to position an aqueous 300 mg/L Stevia solution in each of the rankings of standards. For this purpose, the MP extract was randomly selected among the Stevia samples. For determination of the concentration of the reference, the arithmetic mean of the positions selected for each attribute was calculated, and the resulting proportion was assigned to the respective concentration of the standard. The resulting concentrations of the reference stimuli in a 5% w/v sucrose solution anchored at 50 mm of the line scale were 114 mg/L of caffeine for bitterness, 280 mg/L of glycyrrhizic acid ammonium salt for licorice, and 11 mg/L of iron sulfate for metallic. Then, subjects were further trained employing these solutions as reference stimuli. Training of the Subjects of the Category Specific Scale. The training of the panelists on the sensory space of the Stevia samples was carried out employing the Feedback Calibration Method (FCM) from Compusense Five software. FCM allows the training time of the panel to be reduced significantly by giving instant feedback on the attribute intensity after performing the assessment of a specific attribute.24 An introductory session was used to present the 100 mm line scale. The training consisted of dedicated training sessions for each attribute. At least three training sessions were conducted for each attribute separately. Training sessions in which all attributes were evaluated for one sample composed of mixtures of the reference stimuli were assessed with FCM. In at least six sessions, the levels below and above the anchor of each attribute were evaluated by the panelists. The following concentrations, in a 5% w/v sucrose solution, were evaluated in the training sessions: for bitterness, 75, 85, 95, 114, 150, and 200 mg/L caffeine; for licorice, 140, 210, 280, 420, 490, and 560 mg/L glycyrrhizic acid; for metallic, 8, 9, 11, 13, and 15 mg/L iron sulfate. The maximum concentration represented the highest intensity on the line scale (100 mm), and a blank represented the lowest intensity on the line scale (0 mm). Additionally, a mixture of these standards was prepared to train the panelists in the assessment of all the attributes simultaneously. In the final step of the training, the panel worked on sessions of consensus, where the panelists calibrated themselves with the mean values achieved by the group after evaluating the different Stevia samples. The training aimed to allow the panelists to differentiate small differences between the Stevia samples, as well as a correct use of the scale. The training sessions were carried out until the standard deviation from the mean of the intensity of each attribute (SD) assessed by the 13 subjects was 60 mg/100 mg of extract. Interestingly, MP contained a wide variety of minor steviol glycosides, up to 24 mg/100 mg (Table 2), resulting from the aqueous extraction and membrane purification processes employed for this sample. The slightly higher than 100% total sweetener content obtained for the three samples might result from inaccuracies of the quantification method. Descriptive Sensory Analysis of the Commercial Stevia Extracts. The results of the sensory evaluation of MP, EM, EC1, and EC2 samples are presented in Table 3. The E

dx.doi.org/10.1021/jf502878k | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

Table 2. Quantification of Major and Minor Steviol Glycosides (Milligrams per 100 mg of Extract) from the Stevia Extractsa sensory propertiesb

origin no.

compound

major glycosides Ra SV subtotal minor glycosides 1 I 2 Rk 3 Rn 4 Rd 5 Rj 6 Rm 13 Rh 14 II 16 Ri 21 Rf 22 Rc 23 DuA 25 Rg 26 III 28 Ru 29 Rb 30 Sb 31 DuB subtotal minor transglycosylated products 7 steviol-(glucosyl)6 8 steviol-(glucosyl)5 9 steviol-(glucosyl)6 10 steviol-(glucosyl)6 11 steviol-(glucosyl)5 12 steviol-(glucosyl)4 15 steviol-(glucosyl)5 17 steviol-(glucosyl)4 20 steviol-(glucosyl)4 (rhamnosyl)1 24 steviol-(glucosyl)4 27 steviol-(glucosyl)3 subtotal total 18 19

a

MP

EM

EC1

67.2(±3.5) 14.9(±0.4) 82.1

2.7(±0.03) 6.3(±0.02) 9.0

84.4(±0.6) 17.9(±0.3) 102.3

0.2(±0.01) 0.2(±0.006) 0.6(±0.04) 2.8(±0.16) 0.3(±0.02) 1.6(±0.09) 0.4(±0.02) 0.7(±0.08) 0.05(±0.001) 1.2(±0.1) 8.4(±0.7) 1.3(±0.09) 2.9(±0.1) 0.5(±0.01) 1.6(±0.1) 0.4(±0.002) nd nd 23.2

nd nd nd 0.2(±0.002) nd 0.9(±0.05) nd nd 3.3(±0.05) 0.3(±0.03) 1.3(±0.06) 0.4(±0.01) 1.7(±0.06) nd 0.3(±0.1) 0.8(±0.001) 0.9(±0.05) 0.4(±0.04) 10.5

nd nd nd nd nd nd nd nd nd nd nd 0.0 105.3

EC2

sweet

bitter

refs

102.6 ndc 102.6

8.3 11.2

0.194 112

4, 33 4, 33

0.3(±0.01) nd nd 0.5(±0.001) nd nd nd nd nd 0.1(±0.01) 0.6(±0.04) n.d nd nd nd 0.8(±0.05) 0.3(±0.002) nd 2.6

0.2(±0.02) nd nd 0.1(±0.005) nd nd nd nd nd 0.3(±0.004) nd n.d nd nd nd 1.6(±0.1) nd nd 2.2

nd nd nd 5.3 nd nd nd nd nd nd 27.8 32.9 nd nd 27.3 18.1 26.8 nd

nd nd nd 162 nd nd nd nd nd nd 49 49 nd nd 61 137 84 nd

27 5 5 4, 33 5 5 5 19 5 33, 34 4, 33 4, 33 5 26 4, 33 4, 33 4, 33 5

2.2(±0.03) 5.2(±0.2) 1.7(±0.2) 2.0(±0.6) 5.4(±0.2) 10.2(±0.3) 4.9(±0.08) 11.2(±0.1) 0.9(±0.05)

nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd nd nd nd

nd nd nd nd nd nd nd nd nd

35 35 35 35 35 35 35 35 35

1.3(±0.08) 1.7(±0.07) 46.7 66.2

nd nd 0.0 104.9

nd nd 0.0 104.8

nd nd

nd nd

35 35

The standard deviation (n = 3) is given in parentheses. bTaste threshold (μM). cnd, not detected.

Table 3. Intensity of Sensory Attributes of Stevia Extracts Determined by a Trained Panel (n = 13 in Duplicate)a attribute sweetness bitterness licorice metallic

MP 49.2 59.1 68.3 53.0

± ± ± ±

7.3a 12.6c 12.8c 12.0a

EM 49.5 53.3 58.9 49.1

± ± ± ±

EC1

5.0a 11.1bc 12.3b 7.9a

48.3 49.7 51.2 47.9

± ± ± ±

5.2a 15.2ab 17.9ab 8.1a

EC2 48.0 44.3 48.6 45.3

± ± ± ±

2.4a 16.4a 16.2a 12.6a

a A category specific line scale 0−100 mm was employed for the sensory evaluation. Means and standard deviations are shown (mean ± SD). The same letter indicates that means of intensity are not significantly different between samples at p < 0.05 confidence level.

were perceived as iso-sweet and compared in sweetness with the reference stimulus (5% w/v sucrose solution anchored at 50 mm of the scale), matching the level of sweetness. The evaluation of the 5% w/v sucrose reference yielded 49.8 ± 0.3 in the attribute sweetness, 4.4 ± 7.7 in bitterness, 1.8 ± 2.9 in licorice, and 3.6 ± 8.9 in metallic. In the case of MP, the presence of natural minor glycosides, and eventually the remaining impurities that were not

means of intensity, together with the standard deviations, are reported. An ANOVA was performed on all attributes. The intensity of the attributes bitterness [F(3,36) = 4.19, p = 0.01] and licorice [F(3,36) = 8.17, p = 0.00] differed significantly between samples. On the contrary, the intensity of the attribute metallic did not differ significantly between samples [F(3,36) = 2.39, p = 0.09]. Sweetness was not significantly different between samples [F(3,36) = 0.32, p = 0.814]. All Stevia samples F

dx.doi.org/10.1021/jf502878k | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

Funding

analytically identified in this study, might be responsible for the significantly higher licorice and bitterness intensities. It is worth mentioning that the MP sample has the highest concentration of rebaudioside D, a steviol glycoside containing the highest amount of β-glycosyl moieties, recently characterized by its high sweetness intensity and low bitterness intensity.4 However, the amount of rebaudioside D was below its threshold concentration (Table 2) and, therefore, its contribution to the whole sweetness of the sample might be limited. The additive impact of the different constituents, however, should be considered in further research. MP sample also has the highest amount of rebaudioside C, but still below the threshold for bitterness; again, an additive effect among the different rebaudiosides could contribute to the higher perceived bitterness level and should be considered. EC1 and EC2 samples showed no significant differences for any of the attributes, demonstrating that the intensities of the off-flavor are equally intense when compared under iso-sweet conditions, despite their different contents in rebaudioside A and stevioside (Table 2). This difference in composition, however, has a cost impact that is worth considering for industrial applications. All samples do not significantly differ in metallic taste. The impact of the fraction in the EM sample that remained compositionally unknown (ca. 35% w/w) on the sensory properties could not be assessed. On the other hand, in the EM sample, the concentration of Ra and Sv represents only 9% w/ w of the total composition; instead, transglycosylated constituents (Table 2) accounting for almost 47% w/w are accountable for the remaining sweetness.11 In conclusion, this study shows that the MP sample contains a wide range of natural steviol glycosides. However, when compared at iso-sweet conditions, this Stevia product shows the strongest intensities of the off-taste attributes bitterness, licorice, and metallic compared to Stevia samples obtained by other processing techniques. Furthermore, extracts resulting from similar, solvent-extracted, purification processing but with different composition (EC1 and EC2) present similar sensory properties, although the extracts vary considerably in chemical composition. Finally, considering that Stevia leaves are particularly rich in Sv, a compound with high bitterness offtaste, our results show that the enzymatic modification occurring in the EM sample is an effective alternative to reduce off-tastes; the similar levels in all attributes of EM and EC1 support this conjecture. The UHPLC-MSn method proposed here should allow identifying major and minor steviol glycosides in a single run and could be applied as a routine quality control method for natural and transglycosylated commercial extracts of Stevia.

■

This research was funded by the Chilean Economic Development Agency, CORFO, through the project “Technological platform for formulating naturally sweetened healthy foods”. This project is part of Wageningen UR Chile, the International Center of Excellence for Foods in Chile, Project 11 CEII-9568. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS ́ We acknowledge Romina Montealegre and Ricardo Rodriguez, from the Centro de Aromas y Sabores, DICTUC, for their excellent technical skills in managing the trained panel during the sensory evaluation sessions. We are particularly grateful to Javier Sainz, general manager of Prodalysa, Chile, for very fruitful discussions and permanent collaboration.

■

(1) Brandle, J. E.; Starratt, A. N.; Gijzen, M. Stevia rebaudiana: its agricultural, biological, and chemical properties. Can. J. Plant Sci. 1998, 78, 527−536. (2) Food and Drug Administration. Notification 252 Merisant GRAS Notification for Rebaudioside A. (3) The European Comission. Commission Regulation (EU) No. 1131/2011 of 11 November 2011 Amending Annex II to Regulation (EC) No. 1333/2008 of the European Parliament and of the Council with Regard to Steviol Glycosides. Off. J. Eur. Union 2011 (Dec 11), 205−2011. (4) Hellfritsch, C.; Brockhoff, A.; Stähler, F.; Meyerhof, W.; Hofmann, T. Human psychometric and taste receptor responses to steviol glycosides. J. Agric. Food Chem. 2012, 60, 6782−6793. (5) Ohta, M.; Sasa, S.; Inoue, A.; Tamai, T.; Fujita, I.; Morita, K.; Matsuura, F. Characterization of novel steviol glycosides from leaves of Stevia rebaudiana Morita. J. Appl. Glycosci. 2010, 57, 199−209. (6) Chaturvedula, V. S. P.; Prakash, I. Cucurbitane glycosides from Siraitia grosvenorii. J. Carbohydr. Chem. 2011, 30, 16−26. (7) Chaturvedula, V. S. P.; Prakash, I. Structural characterization and hydrolysis studies of rebaudioside E, a minor sweet component of Stevia rebaudiana. Eur. Chem. Bull. 2013, 2, 298−302. (8) Mizutani, K. T. O. Use of Stevia rebaudiana sweeteners in Japan. In Stevia: the Genus Stevia; Kinghorn, A. D., Ed.; Medicinal and Aromatic Plants − Industrial Profiles; Taylor & Francis: London, UK, 2002; pp 178−195. (9) Joint FAO/WHO Expert Committee on Food Additives (JECFA). Steviol glycosides. Compend. Food Addit. Specif. 2004, 47−50. (10) Yang, M.; Hua, J.; Qin, L. High-Purity Rebaudioside A and Method of Extracting Same; Google Patents, 2011. (11) Abelyan, V. A.; Balayan, A. M.; Ghochikyan, V. T.; Markosyan, A. A. Transglycosylation of stevioside by cyclodextrin glucanotransferases of various groups of microorganisms. Appl. Biochem. Microbiol. 2004, 40, 129−134. (12) Li, S.; Li, W.; Xiao, Q.; Xia, Y. Transglycosylation of stevioside to improve the edulcorant quality by lower substitution using cornstarch hydrolyzate and CGTase. Food Chem. 2013, 138, 2064− 2069. (13) Ye, F.; Yang, R.; Hua, X.; Shen, Q.; Zhao, W.; Zhang, W. Modification of stevioside using transglucosylation activity of Bacillus amyloliquefaciens α-amylase to reduce its bitter aftertaste. LWT−Food Sci. Technol. 2013, 51, 524−530. (14) Geuns, J. M. C. European Stevia Association. Analysis of steviol glycosides. In Stevia and Steviol Glycosides: Properties, Techniques, Uses, Exposure, Toxicology, Pharmacological Effects; Euprint: Heverlee, Belgium, 2010; pp 11−17. (15) Cacciola, F.; Delmonte, P.; Jaworska, K.; Dugo, P.; Mondello, L.; Rader, J. I. Employing ultra high pressure liquid chromatography as the second dimension in a comprehensive two-dimensional system for

ASSOCIATED CONTENT

S Supporting Information *

Purity, fragment ions used to obtain calibration curves of all standards except Ra and Sv; slope and linearity coefficients obtained. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*(E.A.) Phone: +56 2 2354 4253. Fax: +56 2 2354 5803. Email: [email protected]. G

dx.doi.org/10.1021/jf502878k | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

analysis of Stevia rebaudiana extracts. J. Chromatogr., A 2011, 1218, 2012−2018. (16) Jamaluddin Ahmed, M.; Smith, R. M. Determination of stevioside by high-performance liquid chromatography with pulsed amperometric detection. J. Sep. Sci. 2002, 25, 170−172. (17) Clos, J. F.; DuBois, G. E.; Prakash, I. Photostability of rebaudioside A and stevioside in beverages. J. Agric. Food Chem. 2008, 56, 8507−8513. (18) Zimmermann, B. F.; Woelwer-Rieck, U.; Papagiannopoulos, M. Separation of steviol glycosides by hydrophilic liquid interaction chromatography. Food Anal. Methods 2012, 5, 266−271. (19) Chaturvedula, V. S. P.; Rhea, J.; Milanowski, D.; Mocek, U.; Prakash, I. Two minor diterpene glycosides from the leaves of Stevia rebaudiana. Nat. Prod. Commun. 2011, 6, 175−178. (20) ISO 8586:2012. Sensory Analysis  General Guidelines for the Selection, Training and Monitoring of Selected Assessors and Expert Sensory Assessors, 2012. (21) Kemp, S. E. Sensory Evaluation: A Practical Handbook; WileyBlackwell: Chichester, UK, 2009. (22) DuBois, G. E.; Walters, D. E.; Schiffman, S. S.; Warwick, Z. S.; Booth, B. J.; Pecore, S. D.; Gibes, K.; Carr, B. T.; Brands, L. M. Concentrationresponse relationships of sweeteners: a systematic study. In Sweeteners; Walters, D. E., Orthoefer, F. T., DuBois, G. E., Eds.; American Chemical Society: Washington, DC, USA, 1991; Vol. 450, pp 261−276. (23) Schiffman, S. S.; Sattely-Miller, E. A.; Bishay, I. E. Time to maximum sweetness intensity of binary and ternary blends of sweeteners. Food Qual. Pref. 2007, 18, 405−415. (24) Findlay, C. J.; Castura, J. C.; Lesschaeve, I. Feedback calibration: a training method for descriptive panels. Food Qual. Pref. 2007, 18, 321−328. (25) Meilgaard, M. Sensory Evaluation Techniques, 3rd ed.; CRC Press: Boca Raton, FL, USA, 1999. (26) Chaturvedula, V. S. P.; Prakash, I. Structures of the novel diterpene glycosides from Stevia rebaudiana. Carbohydr. Res. 2011, 346, 1057−1060. (27) Chaturvedula, V. S. P.; Clos, J. F.; Prakash, I. Fluorescent light exposure of rebaudioside A in mock beverages under International Conference on Harmonization (ICH) guidelines. Int. J. Chem. 2012, 4. (28) Ohtani, K.; Aikawa, Y.; Kasai, R.; Chou, W.-H.; Yamasaki, K.; Tanaka, O. Minor diterpene glycosides from sweet leaves of Rubus suavissimus. Phytochemistry 1992, 31, 1553−1559. (29) DuBois, G. E.; Stephenson, R. A. Diterpenoid sweeteners. Synthesis and sensory evaluation of stevioside analogs with improved organoleptic properties. J. Med. Chem. 1985, 28, 93−98. (30) Kochikyan, V. T.; Markosyan, A. A.; Abelyan, L. A.; Balayan, A. M.; Abelyan, V. A. Combined enzymatic modification of stevioside and rebaudioside A. Appl. Biochem. Microbiol. 2006, 42, 31−37. (31) Jaitak, V.; Kaul, V. K.; Bandna; Kumar, N.; Singh, B.; Savergave, L. S.; Jogdand, V. V.; Nene, S. Simple and efficient enzymatic transglycosylation of stevioside by β-cyclodextrin glucanotransferase from Bacillus f irmus. Biotechnol. Lett. 2009, 31, 1415−1420. (32) Zimmermann, B. F. Tandem mass spectrometric fragmentation patterns of known and new steviol glycosides with structure proposals: fragmentation patterns of steviol glycosides. Rapid Commun. Mass Spectrom. 2011, 25, 1575−1582. (33) Joint FAO/WHO Expert Committee on Food Additives (JECFA). Steviol Glycosides. Compend. Food Addit. Specif. 73th Meet. FAO JECFA Monogr. 2010, 10, 17−22. (34) Starratt, A. N.; Kirby, C. W.; Pocs, R.; Brandle, J. E.; Rebaudioside, F. A diterpene glycoside from Stevia rebaudiana. Phytochemistry 2002, 59, 367−370. (35) Fukunaga, Y.; Miyata, T.; Nakayasu, N.; Mizutani, K.; Kasai, R.; Tanaka, O. Enzymic transglucosylation products of stevioside: separation and sweetness-evaluation. Agric. Biol. Chem. 1989, 53, 1603−1607.

H

dx.doi.org/10.1021/jf502878k | J. Agric. Food Chem. XXXX, XXX, XXX−XXX