Food Chemistry 141 (2013) 745–753

Contents lists available at SciVerse ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Analytical Methods

Determination of six steviol glycosides of Stevia rebaudiana (Bertoni)from different geographical origin by LC–ESI–MS/MS

0308-8146/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.foodchem.2013.03.041

⇑ Corresponding author. Tel.: +39 089 969252.E-mail address: pmontoro@unisa.it (P. Montoro).

1 These first two authors contributed equally to this study.

Paola Montoro a,⇑,1, Ilaria Molfetta b,1, Mariateresa Maldini a, Lucia Ceccarini b, Sonia Piacente a,Cosimo Pizza a, Mario Macchia b

a Dipartimento di Scienze Farmaceutiche, Università degli Studi di Salerno, Via Ponte don Melillo, 84084 Fisciano (SA), Italyb Dipartimento di Agronomia e Gestione dell’Agroecosistema, Università di Pisa, Via S. Michele degli Scalzi 2, 56124 Pisa, Italy

a r t i c l e i n f o a b s t r a c t

Article history:Received 6 June 2011Received in revised form 25 January 2013Accepted 13 March 2013Available online 21 March 2013

Keywords:Stevia rebaudianaSteviol glycosidesQuantitative analyzesLC–ESI/MS/MS

Liquid chromatography electro-spray tandem mass spectrometry (LC–ESI/MS/MS) was applied to thedetermination of sweet glycosides (steviol glycosides), and toxic aglycon steviol in 24 samples of Steviarebaudiana (Bertoni) aerial parts, which had been experimentally cultivated in Italy, although derivedfrom seeds of different geographical origin. On the basis of the specific fragmentation of these com-pounds, an LC–MS/MS method was developed with the aim of quantifying analytes in plant material.Although toxic steviol was not detectable in all the samples, the samples with the highest levels of steviolglycosides were identified. Analysis of the different samples revealed that they were good quality sam-ples, quality being directly linked to the presence of sweet glycosides in the plants cultivated in Italy,although there were differences in the content of these compounds according to the origin of the seeds,and in particular, a major concentration of compounds with major sweetness activity and minor toxicitywas found in the population coming from Brazil (for example: sample 10, stevioside content 15.74 ± 2.0%p/p and rebaudioside A content 3.09 ± 0.39% p/p of dried plant). Finally, based on this metabolomic tar-geted approach, the results obtained for the samples were treated by Principal Component Analysis, iden-tifying specific genotypic differences based on the geographic origin of the seeds.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The Genus Stevia, family Asteraceae, is made up of more than150 species, and of these, S. rebaudiana, – commonly known as Ste-via – is the only one with sweet tasting properties, which can beattributed to its diterpenoid glycosides (Tanaka, 1982). S. rebaudi-ana (Bertoni) is a herbaceous plant growing in the wild in Paraguayand Brazil where its leaves have been used as a natural sweetener,by the indigenous population, for hundreds of years. This plant is ofgreat importance today because its leaves are used as a non-nutri-tive, high potency sweetener primarily in Japan, Korea, China andSouth America. In addition to these properties, the water extractof S. rebaudiana has beneficial hypoglycemic (Jeppesen, Gregersen,Alstrupp, & Hermansenn, 2002), and hypotensive effects (Chanet al., 2000) and is a source of antioxidant compounds (Xi, Yamag-uchi, Sato, & Takeuchi, 1998). The sweetness of the leaves is due tothe presence of sweet diterpenoid glycosides characterized by anent-kaurene diterpenoid steviol skeleton (ent-13-hydroxy kaur-16-en-19-oic acid). The principal components are stevioside (6–

10%) and rebaudioside-A (2–4%) while other minor glycosidesstructurally related to stevioside are present up to 1–2% in theleaves (Geuns, 2007; Phillips, 1989). Of these sweet diterpenoidglycosides, the most studied is stevioside (Brandle, 2001; Chatsud-thipong & Muanprasat, 2009). Stevioside is a stable compoundwith a sweet index of 300 and used in the food industry in somecountries, where the use of Stevia derivatives in food and drugs isapproved. Stevioside exhibits neither toxicity not mutagen action,and is a low-caloric substance. Stevioside has an additional thera-peutic value, linked to the possibility of it replacing sugar in dietsand to its ability to stimulate insulin secretion in the pancreas,which could result in this compound being used in the treatmentof diabetes and other disturbances of sugar metabolism (Jeppesenet al., 2003; White, Kramer, Campbell, & Bernstein, 1994). Recently,it has been reported that the extract of the aerial part of S. rebau-diana Bertoni possesses antiviral properties and produces a posi-tive therapeutic effect in the treatment of neuralgia, anemia,rheumatism, eczema, dermatitis, and other pathologies (Dozono,1993; Takahashi et al. 2001). Interest in Stevia has increased in re-cent years, as the need to find new high-potency sweeteners assubstitutes for sucrose to alleviate medical and nutritional con-cerns, has grown.

746 P. Montoro et al. / Food Chemistry 141 (2013) 745–753

Known steviol glycosides include stevioside (St), rebaudioside A(RbA), rebaudioside B (RbB), rebaudioside C (RbC), rebaudioside D(RbD), rebaudioside F (RbF), and steviolbioside (Stb). The composi-tion of purified Stevia extracts ultimately depends on the produc-tion approach used by the manufacturer and on theenvironmental and genetic aspect of the original plant. As is thecase for most secondary metabolites, the glycosides profiles of Ste-via are subjected to considerable variability according to geo-graphic area, state of plant maturity, environmental conditions,harvesting and processing. Since changes in the composition ofthe plant material could affect its sweetness, as well as its thera-peutic and toxic activity, strict quality control is critical to ensurethe effectiveness and safety of this species when used in foodand medicine. Quality control procedures for Stevia sweetenersmust be employed to ensure that the standardization and safetyrequirements, as set by the regulatory agencies are met; theserequirements are important both for extracts and isolated samples.Diverse analytical methods have been reported in previously pub-lished works regarding the separation and quantification of thesecompounds from the leaves of S. rebaudiana Bertoni.

A number of techniques, including liquid chromatography (LC),capillary zone electrophoresis, micellar kinetic capillary electro-phoresis, bi-dimensional UHPLC-UV, HPLC coupled with Massspectrometry and NMR have been applied to determine and quan-tify steviol glycosides in S. rebaudiana samples (Cacciola et al.,2011; Hearn & Subedi, 2009; Jaitak, Gupta, Kaul, & Ahuja, 2008;Kolb, Herrera, Ferreyra, & Uliana, 2001; Pieri, Belancic, Morales, &Stuppner, 2011; Woelwer-Rieck, Lankes, Wawrzun, & Wuest2010). Quality assessment relies on the use of validated analyticalmethods. By far the most popular approach to quantify individualsteviol glycosides is LC in combination with UVor MS detection.The recently revisited Joint FAO/WHO Expert Committee on FoodAdditives (JECFA) method (Joint FAO/WHO Expert Committee onFood Additives) is based on RP-HPLC–UV and enables separationand quantification of nine steviol glycosides.

Recentely Gardana, Scaglianti, and Simonetti (2010) reported aUHPLC–MS method for the analysis and preparation of steviol gly-cosides in Stevia samples. Although they use the potentiality ofTandem mass spectrometry for a preliminary qualitative analysis,they have developed a quantitative method based on the detectionof pseudo-molecular ions.

Nowadays it is well known that HPLC–ESI-triple quadruple(QqQ)-MS allows the quantification of metabolites by means of avery sensitive and selective mass tandem experiment such as Mul-tiple Reaction Monitoring (MRM) (Kitteringham, Jenkins, Lane, Elli-ott, & Park, 2009; Montoro et al., 2011). We have applied thismodern technique to a very relevant analytical subject, the devel-opment of an analytical method for the quantification of diterpe-noid glycosides in S. rebaudiana samples.

Stevia could become an interesting and profitable new crop forthe tropics, for warm areas including temperate areas with hot andrainy summers (as an annual summer crop) and large parts of theMediterranean area, (again as an annual crop during spring and au-tumn or an irrigated perennial crop) (Ramesh, Virendra, & Nima,2006). The best suitable areas in Italy are the northern plains whichhave mild rainy summers, coastal centres which have mild cli-mates, and the south and islands, if irrigation is available to sup-port high productivity. In the present study, a quantitativeevaluation of the secondary metabolites in 24 samples of aerialparts of S. rebaudiana plants, cultivated in Italy (Pisa, Tuscany)but of different origins, has been carried out using LC coupled withelectro-spray mass spectrometry (ESI/MS) and tandem MS (ESI/MS/MS) in order to obtain a metabolite profile of the steviolglucosydes.

In order to achieve this objective, multivariate statistical toolssuch as Principal Component Analysis (PCA) can be employed to

put out similarities and correlations among parameters and sam-ples (Granato, Katayama, & Castro, 2012; Mari, Montoro, Pizza, &Piacente 2012)

For these reasons, based on the metabolomic targeted approach,the results obtained for the biological samples were subjected toPrincipal Component Analysis, resulting in the identification ofspecific genotypic differences based on the geographic origin ofthe seeds.

2. Materials and methods

2.1. Chemicals

Compounds 1–6 (Fig. 1) were obtained by Chromadex (Irvine,CA, US). High purity solvents were used for extraction and pur-chased from Carlo Erba (Milano, Italy). HPLC grade methanol, ace-tonitrile and Tri-fluoroacetic Acid were purchased from J.T. Baker(Baker Mallinckrodt, Phillipsburg, NJ, USA). HPLC grade water(18 mX) was prepared using a Millipore (Bedford, MA, USA)Milli-Q purification system.

2.2. ESI/MS and LC–ESI MS/MS analyzes

2.2.1. ESI/MSESI/MS and ESI–MS/MS data for compounds 1–6 were acquired

using an AB Sciex (Foster City, CA, USA) API2000 triple quadrupleinstrument equipped with an ESI source. In this case the massspectrometer was operated in the positive ion mode with opti-mized conditions: de-clustering potential 100 eV, focusing poten-tial 200 eV, entrance potential 10 eV, collision energy 80 eV, andcollision cell exit potential 15 eV.

2.2.2. LC–ESI MS/MSQuantitative analyzes of compounds 1–6 were performed on an

Agilent (Palo Alto, CA, USA) 1100 HPLC system equipped with aWaters Atlantis RP DC18 column (150 � 2.1 mm, 5 lm i.d.) andcoupled to an AB Sciex (AB Sciex, Foster City, CA, USA) API2000 tri-ple quadruple instrument. The mobile phase was generated by agradient elution programme of phase A (H2O acidified with Tri-flu-oroacetic Acid 0.05%) and phase B (CH3CN acidified with Tri-fluoro-acetic Acid 0.05%). Gradient elution started at 35% of solvent B, to80% of solvent B in 25 min, with a flux of 200 lL min�1. TheAPI2000 mass spectrometer was used in the tandem MS mode withMultiple Reaction Monitoring (MRM). The instrument was oper-ated in the positive ion mode with a de-clustering potential of100 eV, focusing potential of 200 eV, entrance potential of 10 eV,collision energy of 80 eV, collision cell exit potential of 15 eV, ionspray voltage of 5000, and capillary temperature of 300 �C

2.3. Sources of plant material and preparation of extracts

Starting plant materials, were derived from the seeds of threedifferent accessions (strains) received from Brazil, San Paulo, Ara-raquara region (samples 1–10) and Paraguay (samples 11–23)and a local nursery in Italy (sample 24). All the plants were ob-tained by seeding.

The plants under study from Brazil and Paraguay were multi-plied by division heads. All plants were grown at the experimentalcentre in Rottaia, Pisa (latitudes 43�41, longitude 10�23). Theplants were planted in pots containing 2/3 of peat moss (organiccarbon C: 40%, organic nitrogen N: 0.8% organic matter: 80%) and1/3 perlite. Irrigation in the period between May and September2008 was provided by a drip line system with a flow rate of 2 l/h. Fertilization was carried out in June 2008 by adding organic-mineral fertilizer containing 0.5% organic nitrogen (N) and 12.5%

Fig. 1. Steviol glycosides from S. rebaudiana.

P. Montoro et al. / Food Chemistry 141 (2013) 745–753 747

of Urea, 5% oxide potassium (K2O) soluble in water and 8.7% carbon(C) of biological origin.

In the first half of June a cut at about 8–10 cm from the base ofthe plant was made to quantify and analyze the biomass.

Plants were extracted with water:ethanol 50:50, for 30 minwith a continuous vortex, at 30 �C, by calculating a plant: solventratio of 1:10. Each extract was diluted 1:100 before being analyzedby means of a chromatographic system.

2.4. Preparation of standards/calibration curves

Stock solutions (l mg mL�1) of the external standards (ES) wereprepared by dissolving each compound in methanol. Stock solutionswere diluted with appropriate amounts of methanol to give solu-tions containing 1, 2, 2.5, 5 or 10 lg mL�1 of each ES. Calibrationcurves were constructed by injecting each standard solution at eachconcentration level in triplicate. The peak areas were calculated and

748 P. Montoro et al. / Food Chemistry 141 (2013) 745–753

plotted against the corresponding concentrations of the standardcompounds using weighted linear regression to generate standardcalibration curves.

2.5. Method validation

The LC–MS/MS method was validated according to the Euro-pean Medicines Agency (EMEA) guidelines relating to the valida-tion of analytical methods (Quality Guidelines: Validation ofanalytical procedures – Text & London, 1995). Precision was eval-uated at four concentration levels for each compound through trip-licate intra-day assays and inter-day assays over 3 days. In allcases, the standard deviation was no higher than ±2.00%. Specific-ity was defined as the non-interference by other analytes detectedin the region of interest. For the LC–MS/MS method, which wasdeveloped on the basis of the characteristic fragmentation of com-pounds 1–6, no other peaks interfered with the analyte in the MS/MS detection mode. Recoveries were estimated through the addi-tion of pre-determined quantities of standard analytes to knownamounts of plant samples. Recovery was calculated according tothe difference between the amount of analyte measured in thespiked sample and the amount of analyte determined in the sam-ple prior to spiking, plus the amount of standard added. The meanrecoveries for the LC–MS and LC–MS/MS methods were deter-mined to be 100 ± 4% and 100 ± 2%, respectively. The calibrationgraphs, obtained by plotting the area of ES against the known con-centration of each compound, were linear in the range used in theanalysis for all analytes. The limit of quantification (LOQ; equiva-lent to the sensitivity of the quantitative method), defined as thelowest concentration of analyte that could be quantified withacceptable accuracy and precision, was estimated by injecting aseries of increasingly diluted standard solutions until the signal-to-noise ratio was reduced to 10. The limit of detection (LOD;equivalent to sensitivity of the qualitative method), defined asthe lowest concentration of analyte that could be detected, wasestimated by injecting a series of increasingly diluted standardsolutions until the signal-to-noise ratio was reduced to 2.

2.6. Qualitative and quantitative analysis of S. rebaudiana samples

For quantitative analysis, a portion of (20 lL) of each diluted(1:100) S. rebaudiana Bertoni aerial parts extract was injected intothe LC–MS/MS system. Quantitative analyzes were replicated fivetimes for each sample extract of Stevia.

2.7. Statistical analysis

Data are expressed as mean ± pooled standard deviation andanalyzed by two-way ANOVA (Kaleidagraph software version 3.6;Synergy Software, Reading, PA).

2.8. Principal Component Analysis (PCA)

PCA using the singular value decomposition method was per-formed for multivariate data analysis, by using the SIMCA-P+ che-mo-metric package (Umetrics AB, Box 7960, SE-907 19 Umeå,Sweden). PCA was performed by applying the measured quantita-tive value of selected markers of the species, and in particular,steviol glycosides Thus, quantitative data of each chemical markerwere used to define a data set with 24 observation and 6 variables.The resulting metabolomics data were processed using SIMCA P+software 12.0 (Umetrix AB, Umea, Sweden) in order to identifysimilarities among herbal samples.

3. Results and discussion

3.1. Morphological analysis

Productivity results, according to the morphological aspects ofthe 24 plants, grouped according to the geographical origin ofthe seeds (Brazil, ten samples, Paraguay, thirteen samples, andItaly, one sample), can be readily summarized. Regarding high pro-ductivity, the best results were shown by plants of Brazilian origin(mean value: 67.51 cm ± 7.43), while less high productivity wasfound in plants from Paraguay, presenting a mean value of62.50 cm (with a DS of 8.66). The third in the group were the Ital-ian type (mean 50.12 cm ± 6.70). The weight of the fresh materialswere, respectively, 117.21 g (DS 10.80), 116.19 g (DS 17.80), and83.04 g (DS 8.24) for materials of Brazilian, Paraguayan and Italianorigin. The highest dried material weight was presented by plantsobtained from Brazilian seeds (21.92 g ± 3.36), followed by plantsobtained from Paraguayan seeds (19.48 g ± 3.66), and finally plantsobtained from Italian seeds (14.11 g ± 2.01). This range of valuessuggests the possibility of producing good yields and income alsoin Italy, for all cultivars from the three different geographicalorigins.

3.2. ESI/MS and ESI–MS/MS analysis of steviol glycosides

Direct flow injection ESI/MS analyzes of standard compounds1–6 (Fig. 1) were performed in the positive ion mode using a massspectrometer equipped with a triple quadruple analyzer (ESI-QqQMS). The ESI/MS/MS spectra of compounds 1–5 obtained in the po-sitive ion mode are displayed in Fig. 2. The ESI/MS spectrum of 1 inthe positive ion mode predominantly showed the pseudo-molecu-lar ion peak [M + H]+ at m/z 827, the MS/MS spectrum of which(Fig. 2.1) exhibited major fragmentation at m/z 665 arising fromthe loss of a glucose unit. The ESI/MS spectrum of 2 in the positiveion mode predominantly showed the pseudo-molecular ion[M + H]+ at m/z 989, the MS/MS spectrum of which (Fig. 2.2) exhib-ited major fragmentation at m/z 827 generated by the loss of a glu-cose unit. The ESI/MS spectrum of 3 in the positive ion modepresented the pseudo-molecular ion peak as a major peak[M + H]+ at m/z 827, the MS/MS spectrum of which (Fig. 2.3)showed major fragmentation at m/z 665 arising from the loss ofa glucose unit. The ESI/MS spectrum of 4 predominantly showedthe pseudo-molecular ion peak [M + H]+ at m/z 665, the MS/MSspectrum of which (Fig. 2.4) exhibited major fragmentation at m/z 503 due to the loss of a glucose unit. The ESI/MS spectrum of 5in the positive ion mode showed the pseudo-molecular ion peak[M + H]+ at m/z 951, the MS/MS spectrum of which (Fig. 2.5) exhib-ited major fragmentation at m/z 805 arising from the loss of arhamnose unit. The ESI/MS spectrum of 6 showed the pseudo-molecular ion peak [M + H]+ at m/z 319, the MS/MS spectrum ofwhich exhibited fragmentation at m/z 275 due to the loss of aCO2 neutral molecule (Fig. 2.6).

3.3. LC/ESI MS/MS analyzes of steviol glycosides and calibration curvesassessment

In order to obtain accurate data concerning the amounts ofsteviol glycosides in the aerial parts of S. rebaudiana plants culti-vated in Italy but produced by seeds derived from different geo-graphical areas, a quantitative LC–ESI/MS/MS method wasdeveloped. For this purpose, preliminary ESI/MS/MS spectra wererecorded following the direct introduction of standards 1–6 intothe ESI source of a Mass Spectrometry instrument equipped witha triple quadruple analyzer. The transition from the specific pseu-do-molecular ion [M + H]+ of 1 to the corresponding product ion

Fig. 2. ESI/MS/MS spectra of compounds 1–6 (steviol glycosides) obtained using a triple quadruple analyzer operating in positive ion mode.

P. Montoro et al. / Food Chemistry 141 (2013) 745–753 749

with the loss of the terminal sugar for stevioside (1), rebaudiosideA (2), rebaudioside B (3), steviolbioside (4), rebaudioside C (5) wasselected to monitor these analytes. On the other hand, fragmenta-tion resulting in the loss of a Carbon anhydride (CO2) unit was cho-sen for steviol (6). The transitions observed during ESI–MS/MSexperiments, were used to develop a selective and sensitive LC–ESI MS/MS method by using the technique of Multiple ReactionMonitoring. Such an experiment uses electro-spray ionization fol-lowed by two stages of mass selection: the first stage (MS1) selectsthe mass of the intact analyte (precursor ion) and, after fragmenta-tion, of the precursor by means of a collision with gas atoms, thesecond stage (MS2) selects a specific product ion (a fragmention) of the precursor, collectively generating a selected reactionmonitoring (plural MRM) assay. The two mass filters produce avery specific and sensitive response for the selected analyte thatcan be used to detect and integrate a peak in a simple one-dimen-sional chromatographic separation of the sample.

The MRM method therefore involved the precursor/producttransitions from m/z 827.0 to m/z 665.0 for compound 1 and 3,from m/z 989.0 to m/z 827.0 for compound 2, precursor/producttransitions from m/z 665.0 to m/z 503.0 for compound 4, from m/z 951.0 to m/z 805.0 for compound 5 and from m/z 319.0 to m/z275.0 compound 6. The calibration curves obtained by plotting

the area of ES (1–6) against the known concentration of com-pounds were linear in the range of 1–10 lg mL�1.

Fig. 3 shows the LC–ESI–MS/MS MRM analyzes of steviol glyco-sides, obtained by analyzing dilutions of the 6 commercial stan-dards. The chromatographic profile contained all the peakscorresponding to the compounds under investigation, with quanti-tative purpose noticeably intensifying.

3.4. Validation of the method

The LC–ESI MS/MS method was validated according to QualityGuidelines: Validation of Analytical Procedures – Text and Meth-odology (ICH Q2), by evaluating the following parameters.

Linearity (calibration curves equations and regression) togetherwith LOQ and LOD for each of the 6 compounds analyzed are re-ported in Table 1.

3.4.1. SensitivityThe sensitivity of the method was determined through LOD and

LOQ. The limit of quantification (LOQ; equivalent to the sensitivityof the quantitative method), is defined as the lowest concentrationof analyte that could be quantified with acceptable accuracy andprecision. The limit of detection (LOD; equivalent to the sensitivity

Fig. 3. Multiple reaction Monitoring (MRM) LC–ESI MS/MS of standard compounds 1–6 for quantitative analysis.

Table 1Quantification and validation data for steviol glycosides, by using MRM LC–MS/MSanalysis.

Calibration curve r2 P-value LOQ (ng/mL) LOD (ng/mL)

1 y = 127e3 x-4094 0.996 <0.01 7.8 ± 0.8 1.8 ± 0.42 y = 70e3 x-1347 0.999 <0.01 5.8 ± 0.5 1.1 ± 0.53 y = 219e3 x-79363 0.992 <0.01 7.4 ± 0.9 1.6 ± 0.34 y = 330e3 x-5810 0.999 <0.01 8.9 ± 0.7 1.9 ± 0.75 y = 35e3 x-5234 0.998 <0.01 3.8 ± 0.1 2.8 ± 0.56 y = 27e3 x-3419 0.999 <0.01 2.1 ± 0.2 0.1 ± 0.3

LOQ and LOD are reported as concentration of the standard compound in theinjected solution.P-value for correlation coefficients.Residuals follow a random distribution.

750 P. Montoro et al. / Food Chemistry 141 (2013) 745–753

of the qualitative method), is defined as the lowest concentrationof analyte that could be detected. LOQ and LOD for compounds1–6 were calculated according to the method proposed by ICHand reported in the experimental section. The values obtainedare reported in Table 1.

3.4.2. SpecificitySpecificity of the method was ascertained by analyzing the

standard and sample solutions. Specificity was confirmed by anal-ysis of standard samples. The LC–MS/MS method, was developedon the basis of the characteristic fragmentation of compounds 1–6, and no other peaks interfered with the analytes in the MS/MSdetection mode.

3.4.3. AccuracyThe recovery test was used to evaluate the accuracy of the ana-

lytical procedure. This involved the addition of known quantities ofthe reference standard compounds, taken from the stock solutionto one of the pre-analyzed samples. Three concentration levelswere tested (low, middle and high) for each of the six compoundsunder investigation. At each level, samples were prepared in tripli-cate and analyzed according to previously described procedure.

3.4.4. PrecisionThe intra-day accuracy and precision were calculated by ana-

lyzing three samples of compound 1 at three different concentra-tion levels, namely, 1, 5 and 10 lg/mL, on the same day. Inter-day estimates were performed over three consecutive days. Thestandard deviation was <5%.

3.4.5. LinearityThe calibration graphs, obtained by plotting the area of an

external standard versus the known concentration of each com-pound, were linear in the range of 1–10 lg mL�1 for all com-pounds, with their regression values reported in Table 1. Thesignificance (P-values) of such correlation was also provided.

3.5. LC–ESI MS/MS analysis of steviol glucosides in S. rebaudiana aerialparts

Owing to the high-polarity of glycosides, the aqueous solutionsof methanol, ethanol or a combination of both were frequentlyused for extraction. In our procedures, plants were extracted with

Table 2Quantitative results for compounds 1–6 in 25 samples of Stevia rebaudiana aerial parts.

Sampl 1 2 3 4 5 6 TOT

1 14.25 ± 0.56 2.87 ± 0.31 n.d. 0.2 ± 0.02 2.45 ± 1.74 n.d. 19.77 ± 0.432 11.26 ± 1.37 2.91 ± 0.32 0.60 ± 0.11 0.59 ± 0.13 2.64 ± 0.27 n.d. 18.00 ± 0.373 12.92 ± 0.50 2.38 ± 0.31 0.08 ± 0.07 0.81 ± 0.15 6.56 ± 1.39 n.d. 22.75 ± 0.404 12.77 ± 0.38 2.98 ± 0.35 n.d. 0.30 ± 0.03 7.39 ± 1.90 n.d. 23.44 ± 0.535 9.60 ± 1.07 3.14 ± 0.39 n.d. 0.71 ± 0.08 2.93 ± 0.73 n.d. 16.38 ± 0.456 12.90 ± 1.58 3.14 ± 0.39 n.d. 0.72 ± 0.08 3.55 ± 0.45 n.d. 20.31 ± 0.507 14.74 ± 1.22 2.36 ± 0.42 0.23 ± 0.07 0.19 ± 0.01 2.64 ± 0.41 n.d. 20.16 ± 0.438 13.63 ± 1.27 3.06 ± 0.37 0.64 ± 0.15 0.26 ± 0.04 2.94 ± 0.64 n.d. 20.63 ± 0.499 13.97 ± 3.03 3.26 ± 0.44 0.23 ± 0.11 0.67 ± 0.08 7.99 ± 0.38 n.d. 26.12 ± 0.8110 15.59 ± 2.09 3.09 ± 0.39 0.03 ± 0.07 0.80 ± 0.10 2.78 ± 0.42 n.d. 22.29 ± 0.5911 15.73 ± 3.20 2.72 ± 0.39 0.64 ± 0.16 1.60 ± 0.30 3.44 ± 0.06 n.d. 24.13 ± 0.8212 14.20 ± 0.38 2.61 ± 0.24 1.31 ± 0.18 1.28 ± 0.17 7.98 ± 0.24 n.d. 27.38 ± 0.2413 8.89 ± 0.80 2.42 ± 0.18 n.d. 0.94 ± 0.18 4.44 ± 0.15 n.d. 16.69 ± 0.3314 12.90 ± 1.00 2.13 ± 0.07 0.80 ± 0.06 1.78 ± 0.35 2.95 ± 0.10 n.d. 20.74 ± 0.3215 9.40 ± 4.74 2.71 ± 0.26 0.02 ± 0.04 0.97 ± 0.15 3.26 ± 0.08 n.d. 16.36 ± 1.0016 4.35 ± 0.92 2.14 ± 0.75 0.50 ± 0.06 0.58 ± 0.10 2.95 ± 0.04 n.d. 10.52 ± 0.4417 4.33 ± 2.36 2.50 ± 0.21 0.32 ± 0.12 0.32 ± 0.04 2.85 ± 0.03 n.d. 10.32 ± 0.5518 6.88 ± 1.70 2.57 ± 0.21 n.d 0.52 ± 0.05 5.88 ± 0.16 n.d. 15.85 ± 0.5319 3.77 ± 0.003 2.37 ± 0.15 0.48 ± 0.12 0.51 ± 0.05 5.07 ± 0.12 n.d. 12.20 ± 0.0920 2.21 ± 0.50 2.37 ± 0.15 0.16 ± 0.06 0.42 ± 0.04 7.56 ± 0.18 n.d. 12.72 ± 0.1921 2.12 ± 0.27 2.22 ± 0.10 0.24 ± 0.07 0.53 ± 0.06 4.33 ± 0.08 n.d. 9.44 ± 0.1222 3.99 ± 0.36 2.24 ± 0.11 0.46 ± 0.09 0.45 ± 0.05 5.77 ± 0.13 n.d. 12.91 ± 0.1723 6.12 ± 1.51 2.29 ± 0.12 0.49 ± 0.08 0.32 ± 0.03 6.45 ± 0.15 n.d. 15.67 ± 0.3824 3.17 ± 0.83 2.38 ± 0.20 0.21 ± 0.03 0.65 ± 0.11 4.99 ± 0.03 n.d. 11.40 ± 0.24

(Value given in% p/p of dried plant).Samples 1–10 = samples obtained from seeds of Brazilian origin.Samples 11–23 = samples obtained from seeds of Paraguayan origin.Sample 24 = samples obtained from seeds of local origin.TOT = Total content (value given in% p/p of dried plant).

P. Montoro et al. / Food Chemistry 141 (2013) 745–753 751

water: ethanol 50:50 (10 mL � 3), for 30 min with a continuousvortex, at 30 �C, by calculating a plant:solvent ratio of 1:10.

Table 2 reports the contents of the 6 compounds in the differentsamples of S. rebaudiana, grown by seeds from different geograph-ical origin.

Fig. 4. Principal Component Analysis Score Plot for samples 1–24 targeted analysis on sgrown from seeds of Paraguayan origin, the green area defines the cluster area where wecultivated locally.

Data were presented as mean ± pooled standard deviation. Amatrix of the data was produced and analyzed and analysis of var-iance, ANOVA, was used for multiple comparison, leading to signif-icant differences between the group’s means.

teviol glycosides. Blue line defines the cluster area where we have mainly sampleshave mainly samples grown from seeds of Brazilian origin. Red samples have been

752 P. Montoro et al. / Food Chemistry 141 (2013) 745–753

Compound 3, rebaudioside B is not detectable in samples 1,4, 5,6 13,18, and is present in low amounts in the other samples. Steviolis not detectable in any of the Stevia samples under investigation.

The samples with a high content of stevioside (1) and rebaudio-side A (2) are samples 8, 9, 10. All these samples are the results ofthe cultivation of seeds of Brazilian origin. This is important infor-mation, because these two glycosides show the highest sweetnessactivity and minor toxicity. In addition, Brazilian origin sampleshave a higher total content of these sweet terpenic glycosides.On the other hand, samples 11 and 12, of Paraguayan origin, arethe only two samples of this origin that showed an high amountof compound 1, but do not show a high level of compound 2. Con-sidering that the total amount of compound 1 and 2 is indicative ofquality as seen in Table 2, it could be said that (with few excep-tions) samples of Brazilian origin can be defined samples of higherquality compared to steviol glycosides.

Principal Component Analysis (PCA), using SIMCA-PLUS Soft-ware, was applied to the simple matrix obtained, as seen in Table 2which shows the quantitative content of each marker compound ineach of the 24 samples. PCA is an unsupervised method and wasused to reduce the dataset in order to obtain the maximum varia-tion between the samples. Unit variance was chosen for scalingdata.

Fig. 4 shows the 2D projection plot of the 24 S. rebaudiana sam-ples. The first component (R1X) explain the 44% of variance whilst,and the second (R2X) the 27%. Principal component’s choice wasdone on the basis of the fitting (R2X) and predictive (Q2X) valuesfor the PCA model, in our case the second component gave the clos-est value to 1 for both of them. Variance was evaluated by Signif-icance level for Hotelling’s T2.

There are confined cluster areas in the 2D diagram, presenting alink to geographical origin; most of the samples generated by seedsof Paraguayan origin are centered in the same area of the graph,and most of the ones relative to the plants grown from seeds ofBrazilian origin, are confined to another area. Blue lines definethe cluster area where we have mainly samples grown from seedsfrom Paraguay, whereas a green area defines the cluster areawhere we have mainly samples grown from seeds of Brazilian ori-gin. Red denotes samples from the local area. Local area samplescan also be found in the area of the samples coming from Para-guayan seeds. The plants originating from Paraguayan seeds haveshown a minor content of stevioside and rebaudioside A. However,this-research has ascertained that good quality plants of Brazilianorigin can be cultivated successfully in Italy. Thus, these seedsseem to be a popular choice for the cultivation of S. rebaudiana inItaly.

4. Conclusions

In the present study an LC–ESI MS/MS method, based on a Mul-tiple Reaction Monitoring technique, was developed for the analy-sis of stevioside (1), rebaudioside A (2), rebaudioside B (3),rebaudioside C (4), steviolbioside (5) and steviol (6) in S. rebaudi-ana Bertoni extracts. The developed method was validated accord-ing to ICH, and found to be accurate, selective and precise in theapplied range of concentration. The LC–ESI/MS/MS method devel-oped was applied to the analysis of real samples, and in particularto extracts obtained by aerial parts of 24 vegetal samples obtainedby the cultivation of S. rebaudiana seeds of different geographicalorigin. The method was specific and sensitive for the analytes stud-ied. Application of this method is suitable for the quality control ofboth crops and products of S. rebaudiana.

Analysis of the different samples revealed good quality levelsamples of the plants cultivated in Italy, as far as the presence ofsweet glycosides was concerned. However, differences in the

content of these compounds were found, depending on the originof the seeds and a major concentration of compounds with majorsweetness activity and minor toxicity, was found in the populationcoming from Brazil. Absence of steviol in all of the analyzed sam-ples, could be an indicator of the absence of primary toxicity inthese samples, when extracted with ethanol:water mixtures. Thesefindings are confirmed by a Principal Component Analysis (PCA).

Liquid chromatography coupled with mass spectrometry (LC–MS) appears to be the ideal method for PCA analysis of food mate-rial. For this reason, approaches based on LC–MS or LC–MS/MStechniques followed by PCA can find useful application in the fieldof qualitative and quantitative analysis of foods.

References

Brandle, J. U. S. (2001). Patent No. 6255557. <http://www.chemweb.com/databases/patents>.

Cacciola, F., Delmonte, P., Jaworska, K., Dugo, P., Mondello, L., & Rader, J. I. (2011).Employing ultra high pressure liquid chromatography as the second dimensionin a comprehensive two-dimensional system for analysis of Stevia rebaudianaextracts. Journal of Chromatography, A, 1218(15), 2012–2018.

Chan, P., Tomlinson, B., Chen, Y., Liu, J., Hsieh, M., & Cheng, J. (2000). A double-blindplacebo-controlled study of the effectiveness and tolerability of oral steviosidein human hypertension. British Journal of Clinical Pharmacology, 50(4), 215–220.

Chatsudthipong, V., & Muanprasat, C. (2009). Stevioside and related compounds:Therapeutic benefits beyond sweetness. Pharmacology and Therapeutics, 121(1),41–54.

Dozono, F. (1993). US Patent No. 5262161. <http://www.chemweb.com/databases/patents>.

Gardana, C., Scaglianti, M., & Simonetti, P. (2010). Evaluation of steviol and itsglycosides in Stevia rebaudiana leaves and commercial sweetener by ultra-high-performance liquid chromatography-mass spectrometry. Journal ofChromatography, A, 1217(9), 1463–1470.

Geuns, J. M. C. (2007) Stevia and steviol glycosides. Properties, techniques, uses,exposure, toxicology, pharmacological effects, EUPRINT.

Granato, D., Katayama, F. C. U., & Castro, I. A. (2012). Characterization of red winesfrom South America based on sensory properties and antioxidant activity.Journal of the Science of Food and Agriculture, 92, 526–533.

Hearn, L. K., & Subedi, P. P. (2009). Determining levels of steviol glycosides in theleaves of Stevia rebaudiana by near infrared reflectance spectroscopy. Journal ofFood Composition and Analysis, 22(2), 165–168.

Quality Guidelines: Validation of analytical procedures – Text and methodology(ICH Q2), London, European Medicines Agency, 1995. Available at: http://www.emea.europa.eu/pdfs/human/ich/038195en.pdf.

Jaitak, V., Gupta, A. P., Kaul, V. K., & Ahuja, P. S. (2008). Validated high-performancethin-layer chromatography method for steviol glycosides in Stevia rebaudiana.Journal of Pharmaceutical and Biomedical Analysis, 47, 790–794.

Jeppesen, P. B., Gregersen, S., Alstrupp, K. K., & Hermansenn, K. (2002). Steviosideinduces antihyperglycaemic, insulinotropic and glucagonostatic effects in vivo:studies in the diabetic Goto-Kakizaki (GK) rats. Phytomedicine, 9(1), 9–14.

Jeppesen, P. B., Gregersen, S., Rolfsen, S. E. D., Jepsen, M., Colombo, M., Agger, A.,et al. (2003). Antihyperglycemic and blood pressure-reducing effects ofstevioside in the diabetic Goto-Kakizaki rat. Metabolism, Clinical andExperimental, 52(3), 372–378.

Joint FAO/WHO Expert Committee on Food Additives (JECFA), Compendium of foodadditive specifications. Monograph 10. Steviol glycosides. Available at: http://www.fao.org/ag/agn/jecfaadditives/specs/monograph10/additive-442-m10.pdf.

Kitteringham, N. R., Jenkins, R. E., Lane, C. S., Elliott, Vi. L., & Park, B. K. (2009).Multiple reaction monitoring for quantitative biomarker analysis in proteomicsand metabolomics. Journal of Chromatography, B: Analytical Technologies in theBiomedical and Life Sciences, 877(13), 1229–1239.

Kolb, N., Herrera, J. L., Ferreyra, D. J., & Uliana, R. F. (2001). Analysis of sweetditerpene glycosides from Stevia rebaudiana: Improved HPLC method. Journal ofAgricultural and Food Chemistry, 49(10), 4538–4541.

Mari, A., Montoro, P., Pizza, C., & Piacente, S. (2012). Liquid chromatography tandemmass spectrometry determination of chimica markers and principal componentanalysis of Vitex agnus-castus L. fruits (Verbenaceae) and derived foodsupplements. Journal of Pharmaceutical and Biomedical Analysis, 70, 224–230.

Montoro, P., Maldini, M., Russo, M., Postorino, S., Piacente, S., & Pizza, C. (2011).Metabolic profiling of roots of liquorice (Glycyrrhiza glabra) from differentgeographical areas by ESI/MS/MS and determination of major metabolites byLC–ESI/MS and LC–ESI/MS/MS. Journal of Pharmaceutical and Biomedical Analysis,54(3), 535–544.

Phillips, K. C. (1989). In T. H. Grenby (Ed.), Developments in sweeteners (3th ed.),Elsevier: London, pp. 1–43.

Pieri, V., Belancic, A., Morales, S., & Stuppner, H. (2011). Identification andquantification of major steviol glycosides in Stevia rebaudiana purifiedextracts by 1H NMR spectroscopy. Journal of Agricultural and Food Chemistry,59(9), 4378–4384.

P. Montoro et al. / Food Chemistry 141 (2013) 745–753 753

Ramesh, K., Virendra, S., & Nima, M. W. (2006). Cultivation of Stevia (Steviarebaudiana (Bert.) Bertoni): A comprehensive review. Advance in Agronomy, 89,137–177.

Takahashi, K., Matsuda, M., Ohashi, K., Taniguchi, K., Nakagomi, O., Abe, Y., et al.(2001). Analysis of anti-rotavirus activity of extract from Stevia rebaudiana.Antiviral Research, 49(1), 15–24.

Tanaka, O. (1982). Steviol-glycosides: New natural sweeteners. TrAC, Trends inAnalytical Chemistry, 1(11), 246–248.

White, J. R., Kramer, J., Campbell, R. K., & Bernstein, R. (1994). Oral use of atopical preparation containing an extract of Stevia rebaudiana and the

chrysanthemum flower in the management of hyperglycemia. Diabetes-Care,17(8), 940–941.

Woelwer-Rieck, U., Lankes, C., Wawrzun, A., & Wuest, M. (2010). Improved HPLCmethod for the evaluation of the major steviol glycosides in leaves of Steviarebaudiana. European Food Research and Technology, 231(4), 581–588.

Xi, Y., Yamaguchi, T., Sato, M., & Takeuchi, M. (1998). Study on Stevia rebaudiana forantioxidant activity. II. Antioxidant mechanism of Stevia rebaudiana extract andantioxidant activity of inorganic salts. Nippon Shokuhin Kagaku Kaishi, 45(5),317–322.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具