Food Chemistry 241 (2018) 452–459

Contents lists available at ScienceDirect

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

Pretreatment with ethanol as an alternative to improve steviol glycosides T

extraction and purification from a new variety of stevia
⁎
Maysa Formigonia, , Paula Gimenez Milania,b, Alexandre da Silva Avíncolac,
Vanessa Jorge dos Santosc, Livia Benossia, Antônio Sergio Dacomeb, Eduardo Jorge Pilauc,
Silvio Claudio da Costaa,b
a
Programa de pós-graduação em Ciência de Alimentos, Centro de Ciências Agrárias, Universidade Estadual de Maringá, Av. Colombo 5790, CEP 87020-900 Maringá, PR,
Brazil
b
Departamento de Bioquímica, Centro de Ciências Biológicas, Universidade Estadual de Maringá, Av. Colombo 5790, CEP 87020-900 Maringá, PR, Brazil
c
Departamento de Química, Centro de Ciências Exatas, Universidade Estadual de Maringá, Av. Colombo 5790, CEP 87020-900 Maringá, PR, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: Leaves of a new variety of Stevia rebaudiana with a high content of rebaudioside A were pretreated with ethanol.
Stevia rebaudiana The ethanolic extract showed high antioxidant potential and 39 compounds were identified, by UPLC/HRMS,
Pretreatment among them one not yet mentioned in the literature for stevia leaves. From the in natura leaves and pretreated
Yield extraction leaves, the conditions of aqueous extraction of steviol glycosides were investigated using response surface
Byproduct
methodology. The aqueous extracts obtained were purified by ion exchange chromatography techniques and
Bioactive potential
membrane separation methods. The recuperation of steviol glycosides was 4.02 g for pretreated leaves and
2.20 g for in natura leaves. The level of purity was, respectively, 87% and 84.8%. The results obtained de-
monstrate that pretreatment increases the yield and purity level of stevia sweeteners by the use of en-
vironmentally friendly methodologies and the final product presented acceptable sensory characteristics.

1. Introduction conventional techniques and modern techniques (Jentzer, Alignan,

Stevia rebaudiana (Bertoni) is a shrub that belongs to the Asteraceae
family, which is native to Paraguayan regions and is also found in Brazil
and Argentina (Lemus-Mondaca, Vega-Galvez, Zura-Bravo, & Ah-Hen,
2012). Its extracts are rich in steviol glycosides and are used as sub-
stitutes for sucrose (Yadav, Singh, Dhyani, & Ahuja, 2011).
In addition to the sweetening properties derived from steviol gly-
cosides, which mainly include Stevioside, Rebaudioside A and C, stu-
dies have shown that extracts from its leaves also have other metabo-
lites with bioactive potential, such as alkaloids, water soluble
chlorophylls, xanthophylls, derivatives of hydroxycinnamoyl (derived
from caffeine and chlorogenic acid), soluble oligosaccharides, free su-
gars, amino acids, lipids, essential oils and trace elements
(Komissarenko, Derkach, Kovalyov, & Bublik, 1994), which are said to
be responsible for producing therapeutic benefits, such as anti-
hyperglycemic, antihypertensive, anti-inflammatory, antitumor, anti-
diarrheal, diuretic and immunomodulatory benefits
(Chatsudthipong & Muanprasat, 2009).
When assessing the extraction and purification of steviol glycosides
obtained from stevia leaves, we find a large body of literature involving
Vaca-Garcia, Rigal, & Vilarem, 2015; Vanneste, Sotto, Courtin, et al.,
2011; Yildiz-Ozturk, Tag, & Yesil-Celiktas, 2014). The extraction pro-
cedures involve solvents, enzymatic extractions, supercritical fluid
methods, microwaves, ultrasound followed by purification involving
columns, solvent-liquid-liquid extraction, ion exchange, ultra-mem-
branes and nanofiltration, crystallization and fractional distillation.
Despite the remarkable progress in extraction, the production of pro-
ducts with low impurity that are scalable using a minimum of solvents
is still difficult (Rao, Reddy, Ernala, Sridhar, & Ravikumar, 2012)
Aiming to obtain sweeteners with acceptable sensorial character-
istics using the minimal possible solvents during processing, the ob-
jective of this study was to achieve an ethanolic treatment of leaves of
Stevia rebaudiana, to characterize the obtained byproduct and its po-
tential as an additive in food, to investigate the changes generated by
the treatment in the characteristics of the leaf and extraction yield and,
finally, to define a methodology to obtain sweeteners and evaluate the
sensorial characteristics of the final product.

⁎
Corresponding author.
E-mail address: mayformigoni@live.com (M. Formigoni).

http://dx.doi.org/10.1016/j.foodchem.2017.09.022
Received 22 March 2017; Received in revised form 24 July 2017; Accepted 6 September 2017
Available online 07 September 2017
0308-8146/ © 2017 Elsevier Ltd. All rights reserved.
M. Formigoni et al.
2. Methods
2.1. Sample and reagents
Leaves of Stevia rebaudiana were obtained at the experimental site
for the Nucleus of Research in Natural Products (NEPRON) study lo-
cated at the State University of Maringá. The leaves of the seminal
variety, Stevia UEM-13, were in the flower bud formation stage (ap-
proximately 50–60 days after pruning) when the glycosides content of
steviol was at a maximum. Afterward, they were dried in a forced cir-
culation air oven until the humidity reached levels below 10%, crushed
and stored.
All solvents and standards were LC grade or higher. Absolute
ethanol (99.9%) by Merck (Londrina, Paraná, Brasil). Deionized water
(18 MΩ·cm) by Milli-Q plus system was purchased from Induslab
(Londrina, Paraná, Brasil). All the reference compounds were provided
by Sigma-Aldrich (Brasil).
2.2. Ethanolic pretreatment
The ethanolic treatment of Stevia rebaudiana leaves was performed
by means of a column. Approximately 300 g of the leaves were packed
with absolute ethanol. Afterward, the same solvent was eluted using
gravity and 14 ethanolic fractions of 350 ml were obtained. Afterward,
the leaves and fractions were dried and stored for subsequently ob-
taining the sweeteners and characterizing the fractions.
2.3. Characterization of leaves
To investigate whether the ethanolic treatment generated changes
in leaf characteristics, the composition was analyzed before and after
treatment.
2.3.1. Physico-chemical analysis
Moisture and ashes were measured according to the methodology
proposed by the Adolfo Lutz Institute (2005). For Chlorophyll A and B,
the Arnon method (1949) was used, and for anthocyanins, the method
described by Lee and Francis (1972) was used.
2.3.2. Fatty acid extraction
Stevia rebaudiana leaves were extracted to obtain total fatty acids
with low toxicity solvents as described by Biondo et al. (2015) with
some adaptations. Ethanol:hexane:water (5:5:2) was used as the sol-
vent, and after separation of phases with a separation funnel, the
hexane phase was dried to obtain the oil and for subsequent char-
acterization of the fatty acids.
2.3.3. GC–MS
For the quantification of fatty acids extracted from stevia leaves,
esterification was performed as described by Santos Júnior et al.
(2014). The fatty acids methyl esters (FAME) were separated in a gas
chromatographer, Thermo, Trace Ultra, equipped with a flame ioniza-
tion detector and fused silica capillary column CP – 7420 (Select FAME,
100 m in length, 0.25 mm in internal diameter, and 0.25 μm of cya-
nopropyl). The exhaust gas flows were 1.2 ml min−1 for the carrier gas
(H2), 30 ml min−1 for the make-up gas (N2), 35 and 350 ml min−1 for
the H2 and the synthetic air, respectively, for the flame of the detector.
The injected volume was 2.0 μl, using a split sample of 1:80. The
temperatures of the injector and detector were 240 °C. A temperature of
165 °C was used in the column, for 7.00 min, followed by a heating
ramp of 4 °C min−1 until reaching 185 °C, remaining in this condition
for 3 min, followed by a new heating ramp of 6 °C min−1 until the
column reached 235 °C, for 1.67 min, totaling therefore 25 min of
analysis. The times of retention and the FAME peak areas were de-
termined using the software ChromQuest™ 5.0.
453
Food Chemistry 241 (2018) 452–459
2.4. Experimental design
Before obtaining the sweetener as a powder, a study of the condi-
tions of extraction was conducted at the central point. It contained
three variables and two levels, totaling three replicates at the central
point with 11 experiments. Response values are expressed as total ste-
viol glycosides extraction yield (mg/g dry leaf).
To evaluate the total yield of steviol glycoside extraction (Y) as a
function of the three independent variables time (x1), temperature (x2)
and leaf:water ratio (x3).
2.5. Obtaining sweetener powder
After studying the influence of the processing variables on the total
yield of steviol glycosides, the ideal extraction conditions were defined.
A methodology was proposed to obtain the sweetener where the
minimum amount of solvents was used. For this, Stevia rebaudiana
leaves were extracted with water (1:27) w/v at 60 °C and 200 rpm in
three cycles. Then, the crude extract was filtered under a vacuum to
remove the suspended particles. For purification, the filtered crude
extract passed through a UF (10 kDa) and NF (500 Da) membrane with
an ion exchange and adsorption column using as eluant ethanol: water
(85:15) v/v. Finally, the material was dried in a rotary evaporator, and
the purified, powdered sweetener was obtained. At the end of each step
of the process, an aliquot was collected to investigate the loss of com-
pounds.
2.6. Quantification of steviol glycosides by HPLC
The total glycosides present in the leaves, in the ethanolic extracts
obtained in the treatment column, in the fractions removed during each
purification step of the sweetener and in the final powdered product
were identified and quantified by High Performance Liquid
Chromatography (HPLC) coupled to an index detector refraction S:32
with a 5 μm NH2 column 125 × 4.6 mm in size using acetonitrile:
water (80:20) v/v as the mobile phase with HPLC grade.
2.7. Identification of compounds, bioactive compounds and antioxidant
activity in vitro
The ethanolic fractions and aqueous extract were analyzed (1 mg/
ml) to evaluate the bioactive potential and its possible use as a by-
product, as well as for the identification of these compounds. The ali-
quots collected after the aqueous extraction process were also in-
vestigated.
2.7.1. Total phenolic compounds
Total phenolic compounds were determined according to the
method described by Singleton, Orthofer, and Lamuela-Raventos
(1999). The absorbance was measured at 760 nm. The total phenolic
concentration was expressed as mg of gallic acid equivalent (EAQ)/mg
extract.
2.7.2. Flavonoids
The quantification of total flavonoids was performed using the
method described by Jia, Tang, & Wu, 1999. All extracts and fractions
were prepared at a concentration of 1 mg/ml of ethanol, except for the
ethyl acetate fraction (0.5 mg/ml). The absorbance of the samples was
measured at 510 nm. Data were expressed as quercetin equivalents.
2.7.3. Antioxidant actives
The free radical scavenging activity of extracts and aliquots was
measured by the ability to eliminate DPPH radicals (Blios, 1958). The
absorbance was measured at 517 nm, and gallic acid was used as the
reference compound. The results were expressed as percent inhibition
of free radicals.
M. Formigoni et al.
2.7.4. Chemical characterization of stevia extracts by UPLC-HRMS
For analysis of the fractions, aliquots of 80 mg/ml extracts were
analyzed with UHPLC-HRMS using a liquid chromatography system,
Nexera X2, with LC-30AD pump and Shimadzu XR-ODSIII150 × 2 mm
column maintained at 40 °C with a linear gradient of elution using
water (0.1% formic acid) (A) and acetonitrile (0.1% formic acid) (B) as
solvent with LC–MS purity. Chromatographic separation was performed
in 20 min (conditions see Supplementary Table 1).
The mass spectrometer used was the Q-TOF type (Bruker,
Germany), with an electrospray ionization source operated in AutoMS/
MS acquisition mode where the 3 most intense ions of each chroma-
tographic peak were selected with 5 Hz (MS and MS/MS) and equip-
ment tune in the range of m/z 70–1300. The analyses were performed in
positive ionization mode with a capillary voltage of 4.50 kV. The
temperature of the source was 200 °C, and the desolvation gas flow was
8 l·min−1.
The experiments were performed using collision-induced dissocia-
tion (DIC) obtained using a collision energy ramp in the range of
15–40 eV and collision gas pressure of 3.06 and 10–3 mBar in the col-
lision chamber. The ion chromatogram and the obtained MS and MS/
MS spectra were visualized with DataAnalysis 4.1 software, compared
to the literature and analyzed using free access mass spectrometry da-
tabases, such as Massbank, Metlin and Human Metabolome Database.
2.8. Sensorial analysis
Purified sweeteners were analyzed as described by (Chranioti,
Chanioti, & Tzia, 2016; Pangborn, 1963; Ye et al., 2013) with some
adaptations to evaluate the overall acceptance of sweetener obtained in
relation to sucralose, the relative sweetness, and the bitter taste. To
evaluate the relative sweetness, a 2 g/100 ml aqueous solution of su-
crose standard was prepared, and the value corresponding to the con-
centrations of 0.008, 0.0133 and 0.0200 g/100 ml was compared. For
bitter taste threshold, sweetener samples were dissolved in water at
concentrations of 0.040, 0.060 and 0.080 g/100 ml.
The analyses were carried out in individual tasting booths under
normal lighting conditions in the sensory analysis laboratory of the
State University of Maringá. The samples were served in 25 ml aliquots
without odour in plastic cups at room temperature (25 °C). The selected
panelists were composed of 15 tasters, five men and nine women aged
19–60 years. The tasters often ate sweeteners.
2.9. Statistical analysis
All analyses were performed in triplicate. The results are presented
as average values with standard error and were analyzed statistically
using one-way ANOVA and Tukey’s test. Statistical significance was
accepted at a level of p < 0.05 using the statistical program
STATISTICA 8.0.
3. Results and discussion
3.1. Ethanolic pretreatment and characterization of the leaves and extracts
obtained
Table 1 shows the data obtained with ethanolic treatment. The
fractions collected and dried showed different characteristics and
weights and different physical appearance, ranging from shiny powder
to oil. The glycoside content, antioxidant activity, and total phenolic
compounds were measured. It was also noted that during ethanolic
treatment, there was loss of sweetener (9.5%), but considering the in-
itial sweetener content in the leaf (10.5%) and the potential use of
extracts obtained as additives, this decrease was modest.
Grozeva et al. (2015) studied stevia extracts with different solvents
and found EC 50 values between 0.7 and 1.3 mg/ml in the ethanolic
extract (95%) for antioxidant activity. Yildiz-Ozturk et al. (2014)
454
Food Chemistry 241 (2018) 452–459
evaluated phenolic compounds in ethanolic extracts and obtained va-
lues between 80.13 and 86.47 μgGAE/mg. The values expressed in the
cited studies were similar to those in the present study. It can be seen
that the antioxidant capacity of the fractions with oily physical char-
acteristics (5, 8, 9, 11) was lower than the other fractions, which is in
agreement with what Muanda, Soulimani, Diop, and Dicko (2011)
found when evaluating aqueous extracts, methanolic extracts, and oil
from leaves of Stevia rebaudiana. After detecting the high bioactive
potential of the extracts, the composition was investigated through
UPLC–HRMS analysis for possible identification of the compounds. The
results obtained are shown in Table 2.
Databases were used to identify the substances present in the
ethanolic fractions, which resulted in the identification of 39 com-
pounds in the positive ionization module ([M+H]) with an accuracy
error of at most 4 ppm. The compounds identified were in agreement
with literature data (Gaweł-Bęben et al., 2015; Karaköse,
Müller, & Kuhnert, 2015; Molina-Calle, Priego-Capote, & Luque de
Castro, 2017; Shivanna, Naika, Khanum, & Kaul, 2013). As shown in the
chromatograms (see Supplementary Fig. 1), the 14 fractions showed the
same chromatographic profile with several peaks present, which pro-
vides an idea of the diversity of the extract that was confirmed with the
identification of the compounds. Among the families found, phenolic
compounds, terpenoids, amino acids, and retinoids were included. A
compound not yet mentioned in the literature on stevia leaves, the
coniferilic aldehyde, was also identified (see Supplementary Fig. 2).
The presence of a variety of phenolic compounds confirms the an-
tioxidant and total phenolic analyses presented in Table 2. This finding
confirms the possibility of the extract being a byproduct because the
phenolic components, particularly flavonoids, may directly contribute
to the antioxidant properties and are suggested to play a preventive role
against the development of cancer and cardiovascular diseases, pro-
viding imminent potential as a food additive (Shukla, Mehta,
Bajpai, & Shukla, 2009).
López, Pérez, Vinuesa, Zorzettoc, and Abianbdef (2016) studied an
ethanolic extract of Stevia rebaudiana leaves for antioxidant activity and
antiproliferative effects in cervical (HeLa), pancreatic (MiaPaCa-2) and
colonic (HCT116) cancer cells. The extract induced cell death in all
three cell lines and was more active against cervical cancer cells (HeLa).
Together, the results show that the extract acts as an antioxidant, but
this activity is not due to stevioside but to other compounds presents in
the ethanolic extract. Juyal, Bisht, and Singh (2010) studied the effect
of an ethanolic extract of stevia leaves on nephrolithiasis, and con-
cluded that prophylactic and therapeutic treatment reduced renal cal-
culus in rats with improved renal function and cytoprotective effects. In
addition to the biological potential, studies have already evaluated the
effects of the application of stevia in foods beyond its sweetening
power. Ruiz-Ruiz, Moguel-Ordoñez, Matus-Basto, and Segura-Campos
(2015) replaced the sugar in bread with an aqueous extract of stevia
and observed inhibition of the alpha-amylase and glucosidase enzymes
and lower microbial growth during storage in the partially substituted
loaves when compared to the control, showing that the biological
properties of the extract remain after the manufacturing processes.
3.2. Leaf characterization
After the treatment of the leaves, the impact was investigated in the
proximate composition and in other phytochemical constituents of the
leaves. Additionally, the lipid profile was characterized (Table 3) be-
cause some of the fractions obtained in the column had an oily physical
appearance.
Note that there was a significant decrease in treated leaves in almost
all constituents analyzed. This decrease was because the ethanolic
treatment carried these compounds into their fractions. A decrease in
fatty acid levels was also confirmed (approximately 68%). Interestingly,
it was found that the antioxidant activity, phenolic compounds and
flavonoids did not decrease, but increased, and also, the arachidic and
M. Formigoni et al. Food Chemistry 241 (2018) 452–459

Table 1
Characterization of the ethanolic fractions obtained through the treatment column of stevia leaves.

Fraction Physical characteristics Dry weight (g) Glycoside (mg/g of dry extract) Total phenolics (mg/g of dry extract) EC50 (mg/g of dry extract)

Appearance Stevioside Rebaudioside C Rebaudioside A

01 Bright powder 2.610 0.118 0.093 0.149 61.93 0.67

02 Bright powder 4.150 0.274 0.130 0.230 62.41 0.68
03 Oily powder 3.840 0.304 0.146 0.272 64.85 0.67
04 Powder 1.410 0.198 0.103 0.177 105.55 0.77
05 Oil 1.360 0.159 0.082 0.137 82.95 1.06
06 Oily powder 0.930 0.140 0.070 0.116 94.43 0.84
07 Oily powder 0.890 0.133 0.063 0.109 83.57 0.93
08 Oil 0.910 0.114 0.053 0.089 69.59 3.25
09 Oil 0.800 0.107 0.052 0.082 79.12 1.65
10 Oily powder 0.610 0.104 0.047 0.076 82.15 0.86
11 Oil 0.620 0.096 0.042 0.070 94.80 0.86
12 Oily powder 0.660 0.090 0.040 0.059 83.46 0.84
13 Oily powder 0.480 0.074 0.037 0.048 92.53 0.84
14 Oily powder 0.500 0.074 0.035 0.047 87.09 0.85

Table 2
Identification of compounds and main parameters of identification in ethanolic fractions obtained from the Stevia rebaudiana leaf treatment column.

Family Compound Molecular formula Retention time Adduct m/z Error Fragments

Phenolic compounds
Apigenin-7-O-glucoside C21H20O10 9.97 M+H 433.1119 2.30 271
Derivative Catechin C15H14O6 12.09 M+H 291.1869 291, 290
Kaempferol-3-rhamnoside C21H20O10 10.66 M+H 433.1124 1.15 284, 129, 85
Luteolin-7-O-glucoside C21H20O11 9.35 M+H 449.1070 1.78 287, 448
Quercetrine C21H20O10 10.05 M+H 449.1071 1.55 303, 287, 129, 85
Quercetin-3-arabinoside C20H18O11 9.97 M+H 435.0912 2.29 303
Rutine C27H30O16 9.19 M+H 611.1597 1.63 303, 129
Cinnamic acid and derivatives Cinnamic acid C9H8O2 14.55 M+H 149.0531 4.02 149, 122, 94, 80
Rosmarinic acid C18H16O8 12.74 M+H 361.0912 1.66 361
Alcohols and polyols Chlorogenic acid C16H18O9 8.28 M+H 355.1016 2.25 340, 163
3-5-Dicapheoylquinic acid C25H24O12 10.24 M+H 517.1333 1.54 163, 499
3-4-Dicapheoylquinic acid C25H24O12 10.48 M+H 517.1327 2.70 163, 499
Phenylpropanoid Coniferic aldehyde* C10H10O3 11.13 M+H 179.0702 0.55 179, 161, 147, 133, 119

Terpenoids
Diterpenoids Dulcoside A C38H60O17 11.66 M+H 789.3890 1.64 465, 309, 163, 147
Rebaudioside A C44H70O23 10.16 M+H 967.4352 2.99 325, 481
Rebaudioside B C38H60O18 11.40 M+H 805.3837 1.86 319, 325, 481
Rebaudioside C C44H70O22 11.61 M+H 951.4406 2.73 465, 309
Rebaudioside D C50H80O28 10.28 M+H 1129.4877 2.83 325, 163
Rebaudioside E C44H70O23 11.25 M+H 967.4358 2.37 325, 163
Rebaudioside F C43H68O22 11.55 M+H 937.4252 2.45 325, 319, 163
Steviol C20H30O3 10.35 M+H 319.2265 0.93 319, 301, 283
Steviolbioside C32H50O13 11.37 M+H 643.3314 1.55 319, 163, 145
Sesquiterpenoids Sterebin B/C C20H32O5 11.98 M+H 353.2319 1.13 353, 335, 317, 293
Sterebin D C18H30O3 12.69 M+H 295.2262 2.03 277, 259, 219
Sterebin E/F C20H34O4 13.9514.44 M+H 339.2530 339.2500 0 321, 303
Sterebin G C20H34O5 12.05 M+H 355.2474 1.40 337, 319

Other families
Amino acids Leucine/Isoleucine C6H13NO2 1.69 M+H 132.1019 0 86
Valine C5H11NO2 1.28 M+H 118.0863 0 72
Phenylalanine C9H13NO2 7.71 M+H 166.0859 2.40 103
Proline C4H9NO2 1.35 M+H 116.0706 0 70, 115
Serina C3H7NO3 1.18 M+H 106.1499 0 106, 88
Threonine C4H9NO3 1.21 M+H 120.0654 0.83 74, 102, 119
Glutamine C5H19N2O3 1.21 M+H 147.0763 0.67 130, 84
Tryptophan C11H12N2O2 5.98 M+H 205.0968 1.95 188, 146, 119
Retinoids Retinol C20H30O 14.32 M+H 287.2368 0.34 287, 269, 187, 109
Retinal C20H28O 13.24 M+H 285.2211 0.70 285, 267, 187, 119

* Compound identified not reported in other studies.

beenic fatty acids were detected only in the treated leaves. For total
glycosides, values lower than Gardana, Scaglianti, and Simonetti (2010)
(13.8 g/100 g) and superior to Gardana et al. (2010) were found.
However, neither of the two papers used leaves where rebaudioside A
was the major glycoside as in the present study (4.4 g/100 g for un-
treated leaves and 3.9 g/100 for treated leaves). The values obtained in
this study for both untreated and treated leaves were similar to those
reported in the literature (Mishra, Singh, Kumar, & Prakash, 2010;
Shivanna et al., 2013). Muanda et al. (2011) evaluated antioxidant
activity in aqueous extracts of stevia leaves and obtained 82.86% in-
hibition. Gaweł-Bęben et al. (2015) estimated for aqueous extracts of
stevia leaves phenolic compounds and flavonoids, respectively, 3.85 mg
EAG/g from dry leaf and 2.03 mg EQ/g from dry leaf. The values for
antioxidant activity, phenolic compounds and flavonoids were similar

455
M. Formigoni et al. Food Chemistry 241 (2018) 452–459

Table 3 (Jentzer et al., 2015). Although in this experiment the temperature was
Proximal composition and other phytochemical constituents in leaves of Stevia rebaudiana not a variable with a significant influence, Erkucuk, Akgun, and Yesil-
with and without ethanolic treatment.
Celiktas (2009) and Jentzer et al. (2015) found an increase in yield with
Analysis Untreated leaf* Leaf with treatment* increasing temperature due to the higher energy required to break the
analyte-matrix bonds, and thus the diffusion of these analytes was fa-
Total Glycoside (g/100 g) 10.5a ± 0.44 9.5b ± 0.39 cilitated.
Stevioside/Rebaudioside C/ 39/19/42 39/20/41
Rebaudioside A (%)
Total chlorophyll (mg/g) 2.94a ± 0.69 2.1b ± 0.55 3.4. Influence of pretreatment on extraction yield
Anthocyanins (μg/100 g) 0.52a ± 0.78 0.23b ± 0.94
Phenolic compounds (mg EAG/g dry 5.54 ± 0.09 6.72b ± 0.12 With the optimized processing parameters, treated and untreated
leaf)** stevia leaves were extracted. Fig. 2 shows the glycoside extraction
Flavonoids (mg EQ/g dry leaf)** 1.53a ± 0.05 1.68a ± 0.15
yields considering the initial content of sweeteners present in the
Antioxidant activity (% Inhibition)** 88.72a 94.55b
Moisture (%) 5.92a ± 0.2 7.25b ± 0.24 leaves.
Ashes (%) 9.71a ± 0.09 8.86b ± 0.12 The extraction yield of treated leaves was higher than that of un-
Total lipids treated leaves. Considering that there was also an increase in phenolic
Extractions (g AG/100 g lipids) compounds, flavonoids, and antioxidant activity and just two fatty
Myristic acid Nd 0.17 ± 0.48 acids were detected only in treated leaves (Table 3), it was hypothe-
Palmitic acid 38.41a ± 3.17 13.47b ± 2.08 sized that the ethanolic treatment in stevia leaves facilitates the ex-
Stearic acid 8.11a ± 7.58 0.91b ± 0.67
Oleic acid 13.08a ± 12.20 2.11b ± 5.02
traction of compounds.
Linoleic acid 11.05a ± 4.17 2.58b ± 0.25
Alpha-linolenic acid 6.35a ± 7.06 2.90b ± 4.32 3.5. Purification and obtaining of powdered sweetener
Arachidic acid Nd 0.26 ± 0.31
Eicosanoic acid 9.35a ± 6.97 3.43b ± 0.46
For the use of sweeteners as a food additive, dietary supplement or
Beenic acid Nd 0.34 ± 0.20
Lignoceric acid 7.34a ± 0.22 3.20b ± 1.09 drug, safe and sustainable purification techniques are required.
Currently, the use of solvents, such as ethanol, methanol, and iso-
*
Mean values of triplicate ± SD; **Analyses performed on aqueous extract after extrac- butanol, associated with unit operations is used to achieve purity
tion. Nd: Not detected. Different letters indicate significantly different values (p < 0.05). standards (95%) stipulated by International Scientific Expert
Committee – JECFA (2007), but because it is a product of natural origin
to those in this study. Tadhani and Subhash (2006) investigated fatty with so many known health benefits, the use of toxic solvents should be
acids in stevia leaf oil, and of the 6 types they found, 4 were also ob- avoided at all costs. Fig. 3 shows the proposed flowchart of purification
tained in this study with similar values. The fatty acids known in this of the sweeteners, their losses during processing and final purity.
study represent a greater variety not yet explicit in the literature (see The same methodology for obtaining sweetener from untreated
Supplementary Fig. 3). leaves was used. Starting from the same amount of leaves, the final
mass was 2.20 g with 84.8% purity (see Supplementary Fig. 4), a yield
3.3. Experimental design value of the final product considerably lower when compared to the
same process with treated leaves. Suppose that during the purification
A first order polynomial model that described its behaviour with the of the extract of untreated leaves, the sweeteners were retained along
following polynomial equation was used in the Protimiza Experimental with macromolecules that the ethanolic treatment eluted, justifying the
Design program: lowest yield. Vanneste et al. (2011) studied the performance of a pro-
cess of purification of stevia sweeteners by three stages of membranes
Y= 139.57 + 2.87 + 2.66x2 + 19.37x3−3.31x1x2−7.59x1x3−5.07x2x3
and reached a purity of 37.0% after this stage. Although the final
where the response is a value and x1, x2 and x3 are the coded values of product did not achieve the required purity as stipulated by JECFA
the three independent variables. After the optimization process was (2007) because it is a simple, rapid and scalable methodology without
defined, treated and untreated leaves were extracted to determine if the the use of toxic solvents, a high purity was achieved. To achieve a purer
treatment would influence any steviol glycosides extraction yield. product associated with less loss in the course of unit operations, fur-
After treatment of the leaves, to extract the maximum amount of ther studies must be performed to improve the process
steviol glycosides and to obtain the powdered sweetener, an experiment
was performed using three variables and a response, glycoside extrac- 3.6. Sensorial analysis
tion yield (see Supplementary Table 2).
The effects of different combinations of time, temperature and R-F:S Sucralose is currently the most widely used sweetener in the world
on the extraction yield of the glycosides while maintaining rotation for both industrial and personal use. Although initially considered safe,
(200 rpm) and cycles (3) were investigated. It was evident that time and current studies (Oliveira, Menezes, & Catharino, 2015;
temperature did not influence the extraction of the glycosides when Schiffman & Rother, 2013) warn of the intrinsic biological effects ex-
compared to R-F:S and that the yield increased linearly as the leaf: hibited by sucralose, as well as the potential of its structure to hydro-
solvent ratio increased (Fig. 1). Within the range of R-F:S (1:30), the lyze in toxic compounds when exposed to severe temperature condi-
extraction yield was the highest, which can be explained due to the tions, forming chloropropanes and other related chlorinated
higher amount of solvent extraction and consequent extraction acces- compounds. With this issue in mind, and with the aid of consumers who
sibility. Additionally, when related to time and temperature, these already consume regular sweeteners, the overall acceptance of the
variables were shown to be inversely proportional, where a high yield sweetener obtained in relation to sucralose was evaluated. Additionally,
was present at a higher temperature with shorter time and at a lower to investigate the characteristics of the obtained product better, bit-
temperature in longer time. After data analysis, the optimal extraction terness and relative sweetness were assessed.
parameters of 60 °C (X1), 60 min (X2) and 1:27 w/v (X3) were defined The overall acceptance test involves scores from 1-I greatly disliked
as the optimal extraction parameters. In general, the extraction effi- it up to 9-I liked it a lot. It was observed that there was no significant
ciency of a compound is influenced by multiple parameters, such as the difference in the acceptance of the sweetener consumed by the treated
time, temperature, and polarity of the solvent (type of solvent), and its stevia leaves when compared to sucralose (7.00 ± 1.95 and
effects may be independent or in interaction with other parameters 6.33 ± 1.72, respectively), thus demonstrating that the sweetener was

456
M. Formigoni et al.
Fig. 1. 3D response surface of glycosides showing the effect of the variables of time (X1), temperature (X2) and ratio leaf: solvent – R-F:S (X3), where (A) represents the ratio leaf: solvent –
R-F: S × Time, (B) Temperature × Time and (C) ratio leaf: solvent – R-F: S × Temperature.
accepted.
Relative sweetness was strongly dependent on the equivalence of
sucrose for all high potency sweeteners. It is known that rebaudioside A
sweetens 200 times more than sucrose 6 g/100 ml, but its sweetness
potency is 400 times higher (Schiffman, Booth, Losee,
Pecore, & Warwick, 1995). In this study, the sweetener obtained was
165 times (165.4 ± 0.003) sweeter than sucrose, and the bitterness
threshold was set at 0.073 ± 0.013 g/100 ml. Ye et al. (2013)
457
Food Chemistry 241 (2018) 452–459
Fig. 2. Comparison of glycoside extraction yield (%) with treated and
untreated leaves.
analyzed the relative sweetness and threshold of bitterness of stevioside
and stevioside modified enzymatically with 38.3% conversion, and
obtained values of 173 ± 19.0 and 202 ± 16.2 for sweetness and
0.0172 ± 0.0050 g/100 ml and 0.0263 ± 0.0056 g/100 ml for bit-
terness, respectively. It was observed that the relative sweetness of the
APFT sweetener obtained and evaluated in this study was reduced. This
reduction may have occurred because the purity of the sweetener was
not as high as that used in other work. However, when considering the
M. Formigoni et al.
Fig. 3. Flowchart for obtaining the sweetener powder.
bitterness threshold, the stevia sweetener had a considerably higher
concentration and proved to be less bitter.
4. Conclusion
The treatment of stevia leaves with ethanol before the extraction
process of steviol glycosides is an innovative approach that removes,
selectively, substances such as phenolic compounds and flavonoids,
which may contribute to the residual bitter taste in the final product. In
the extract ethanolic was identified several compounds that present the
potential for use as food additive or medicines. Using a methodology for
obtaining sweeteners, an increase of 43% in yield was obtained from
treated leaves when compared to untreated leaves. Furthermore, sen-
sorial evaluation demonstrated that stevia extract from pretreated
leaves presents sensory profile similar to sucralose.
Acknowledgement
This work was supported by Coordination for the Improvement of
Higher Education Personnel – CAPES.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.foodchem.2017.09.022.
References
Arnon, D. I. (1949). Copperenzymes in isolatedchloroplasts: Polyphenoloxydase in Beta
vulgaris. Plant Physiology, Maryland, 24(1), 1–15.
Biondo, P. B. F., Dos Santos, V. J., Montanher, P. F., Junior, O. O. S., Matsushita, M.,
Almeida, V. C., & Visentainer, J. V. (2015). A new method for lipid extraction using
low-toxicity solvents developed for canola (Brassica napus L.) and soybean (Glycine
max L. Merrill) seeds. Analytical Methods, 7, 9773–9778.
Blios, M. S. (1958). Antioxidant determinations by the use of a stable free radical. Nature,
26, 1199–1200.
Chatsudthipong, V., & Muanprasat, C. (2009). Stevioside and related compounds:
Therapeutic benefits beyond sweetness. Pharmacology & Therapeutics, 121, 41–54.
458
Food Chemistry 241 (2018) 452–459
Chranioti, C., Chanioti, S., & Tzia, C. (2016). Comparison of spray, freeze and oven drying
as a means of reducing bitter aftertaste of steviol glycosides (derived from Stevia
rebaudiana Bertoni plant) – Evaluation of the final products. Food Chemistry, 190,
1151–1158.
Erkucuk, A., Akgun, I. H., & Yesil-Celiktas, O. (2009). Supercritical CO2 extraction of
glycosides from Stevia rebaudiana leaves: Identification and optimization. Journal of
Supercritical Fluids, 5, 35–89.
Gardana, C., Scaglianti, M., & Simonetti, P. (2010). Evaluation of steviol and its glyco-
sides in Stevia rebaudiana leaves and commercial sweetener by ultra-high-perfor-
mance liquid chromatography–mass spectrometry. Journal of Chromatography A,
1217, 1463–1470.
Gaweł-Bęben, K., Bujak, T., Nizioł-Łukaszewska, Z., Antosiewicz, B., Jakubczyk, A., Karaś,
M., & Rybczyńska, K. (2015). Stevia rebaudiana bert. Leaf extracts as a multifunctional
source of natural antioxidants. Molecules, 20, 5468–5486.
Grozeva, N., Pavlov, D., Petkova, N., Ivanov, I., Denev, P., Pavlov, A., ... Dimanova-
Rudolf, M. (2015). Characterisation of extracts from Stevia rebaudiana bertoni leaves.
International Journal of Pharmacognosy and Phytochemical Research, 7(6), 1236–1243.
Instituto Adolfo Lutz (2005). Normas Analíticas do Instituto Adolfo Lutz: Métodos
químicos e físicos para análise de alimentos, v. 1, 4ªed., Cap. IV, IMESP, São Paulo. p.
98–105.
JECFA (2007). Steviol glycosides. In: Combined Compendium of Food Additive
Specifications, 68th Meeting of the Joint FAO/WHO Expert Committee on Food
Additives. Food and Agriculture Organization of the United Nations (FAO), Rome,
Italy, FAO/JECFA Monograph 4. pp. 61–64.
Jentzer, J.-B., Alignan, M., Vaca-Garcia, C., Rigal, L., & Vilarem, G. (2015). Response
surface methodology to optimize accelerated solvent extraction of steviol glycosides
from Stevia rebaudiana Bertoni leaves. Food Chemistry, 166, 561–567.
Jia, Z., Tang, M., & Wu, J. (1999). The determination of flavonoid contentes in mulberry
and their scavenging effects on superoxide radicals. Food Chemistry, 64, 555–559.
Juyal, D., Bisht, G., & Singh, A. (2010). Antilithiatic effect of ethanolic extract of Stevia
rebaudiana Bert. Pharmacologyonline, 2, 517–523.
Karaköse, H., Müller, A., & Kuhnert, N. (2015). Profiling and quantification of phenolics
in Stevia rebaudiana leaves. Journal of Agricultural and Food Chemistry, 63, 9188–9198.
Komissarenko, N. F., Derkach, A. I., Kovalyov, I. P., & Bublik, N. P. (1994). Diterpene
glycosides and phenylpropanoids of Stevia rebaudiana Bertoni. Rastitel’Nye Resursy, 1,
53–64.
Lee, D. H., & Francis, F. J. (1972). Standardization of pigment analyses in cranberries.
HortScience Stanford, 7(1), 83–84.
Lemus-Mondaca, R., Vega-Galvez, A., Zura-Bravo, L., & Ah-Hen, K. (2012). Stevia re-
baudiana Bertoni, source of a high-potency natural sweetener: A comprehensive re-
view on the biochemical, nutritional and functional aspects. Food Chemistry, 132(3),
1121–1132.
López, V., Pérez, S., Vinuesa, A., Zorzettoc, C., & Abianbdef, O. (2016). Stevia rebaudiana
ethanolic extract exerts better antioxidant properties and antiproliferative effects in
tumour cells than its diterpene glycoside stevioside. Food and Function, 2016(7),
2107–2113.
Mishra, P., Singh, R., Kumar, U., & Prakash, V. (2010). Stevia rebaudiana – A magical
sweetener. Global Journal of Biotecnology & Biochemistry, 5, 62–74.
Molina-Calle, M., Priego-Capote, F., & Luque de Castro, M. D. (2017). Characterization of
Stevia by LC-QTOF MS/MS analysis of polar and non-polar extracts. Food Chemistry,
219, 329–338.
Muanda, F. N., Soulimani, R., Diop, B., & Dicko, A. (2011). Study on chemical compo-
sition and biological activities of essential oil and extracts from Stevia rebaudiana
Bertoni leaves. LebensmittelWissenschaft + Technologie / Food Science + Technology,
44, 1865–1872.
Oliveira, D. N., Menezes, M., & Catharino, R. R. (2015). Thermal degradation of sucralose:
A combination of analytical methods to determine stability and chlorinated by-
products. Scientific Reports, 5, 9598.
Pangborn, R. M. (1963). Relative taste intensities of selected sugars and organic acids.
Journal of Food Science, 28(6), 726–733.
Rao, A. B., Reddy, G. R., Ernala, P., Sridhar, S., & Ravikumar, Y. V. L. (2012). An im-
provised process of isolation, purification of steviosides from Stevia rebaudiana
Bertoni leaves and its biological activity. International Journal of Food Science and
Technology, 47, 2554–2560.
Ruiz-Ruiz, J. C., Moguel-Ordoñez, Y. B., Matus-Basto, A. J., & Segura-Campos, M. R.
(2015). Antidiabetic and antioxidant activity of Stevia rebaudiana extracts (Var.
Morita) and their incorporation into a potential functional bread. Journal of Food
Science and Technology, 52(12), 7894–7903.
Santos Júnior, O. O., Montanher, P. F., Bonafé, E. G., Prado, I. N., Maruyama, S. A.,
Matsushita, M., & Visentainer, J. V. (2014). A simple, fast and efficient method for
transesterification of fatty acids in foods assisted by ultrasound energy. Journal of
Brazilian Chemistry Society, 25(9), 1712–1719.
Schiffman, S. S., Booth, B. J., Losee, M. L., Pecore, S. D., & Warwick, Z. S. (1995).
Bitterness of sweeteners as a function of concentration. Brain Research Bulletin, 36,
505–513.
Schiffman, S. S., & Rother, K. I. (2013). Sucralose, a synthetic organochlorine sweetener:
Overview of biological issues. Journal of Toxicology and Environmental Health B,
Critical Reviews, 16, 399–451.
Shivanna, N., Naika, M., Khanum, F., & Kaul, V. K. (2013). Antioxidant, anti-diabetic and
renal protective properties of Stevia rebaudiana. Journal of Diabetes Complications,
27(2), 103–113.
Shukla, S., Mehta, A., Bajpai, V., & Shukla, S. (2009). In vitro antioxidant activity and
total phenolic content of ethanolic leaf extract of Stevia rebaudiana Bert. Food
Chemistry and Toxicology, 4, 2338–2343.
Singleton, V. L., Orthofer, R., & Lamuela-Raventos, R. M. (1999). Analysis of total phenols
and other oxidation sustrates and antioxidants by means folin-ciocalteu reagentes.
M. Formigoni et al.
Methods in Enzymology, 299, 152–178.
Tadhani, M., & Subhash, R. (2006). Preliminary studies on Stevia rebaudiana leaves:
Proximal composition, mineral analysis and phytochemical screening. Journal of
Medical Sciences, 6, 321–326.
Vanneste, J., Sotto, A., Courtin, C. M., et al. (2011). Application of tailor-made mem-
branes in a multi-stage process for the purification of sweeteners from Stevia re-
baudiana. Journal of Food Engineering, 103, 285–293.
Yadav, A. K., Singh, S., Dhyani, D., & Ahuja, P. S. (2011). A review on the improvement of
459
Food Chemistry 241 (2018) 452–459
Stevia rebaudiana (Bertoni). Canadian Journal of Plant Science, 91(1), 1–27.
Ye, F., Yang, R., Hua, X., Shen, Q., Zhao, W., & Zhang, W. (2013). Modification of ste-
vioside using transglucosylation activity of Bacillus amyloliquefaciens amylase to re-
duce its bitter aftertaste. LWT – Food Science and Technology, 51(2), 524–530.
Yildiz-Ozturk, E., Tag, O., & Yesil-Celiktas, O. (2014). Subcritical water extraction of
steviol glycosides from Stevia rebaudiana leaves and characterization of the raffinate
phase. The Journal of Supercritical Fluids, 95, 422–430.