See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/257837886

Isolation by pressurised fluid extraction (PFE)

and identification using CPC and HPLC/ESI/MS
of phenolic compounds...
Article in Food Chemistry February 2014
DOI: 10.1016/j.foodchem.2013.08.065 Source: PubMed

CITATIONS

READS

13

133

4 authors, including:
Alessandra L. Oliveira
University of So Paulo
42 PUBLICATIONS 396 CITATIONS
SEE PROFILE

All content following this page was uploaded by Alessandra L. Oliveira on 21 August 2014.
The user has requested enhancement of the downloaded file. All in-text references underlined in blue
are linked to publications on ResearchGate, letting you access and read them immediately.

Food Chemistry 145 (2014) 522529

Contents lists available at ScienceDirect

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

Isolation by pressurised uid extraction (PFE) and identication using

CPC and HPLC/ESI/MS of phenolic compounds from Brazilian cherry
seeds (Eugenia uniora L.)
Alessandra L. Oliveira a,, Emilie Destandau b, Latitia Fougre b, Michel Lafosse b
a
Departamento de Engenharia de Alimentos, Faculdade de Zootecnia e Engenharia de Alimentos, Universidade de So Paulo, Av. Duque de Caxias Norte, 225, 13635 900 Pirassununga,
SP, Brazil
b
Institut de Chimie Organique et Analytique, Universit dOrlans, CNRS UMR 6005, 2 Rue de Chartres, 45067 Orlans, France

a r t i c l e

i n f o

Article history:
Received 24 April 2013
Received in revised form 6 August 2013
Accepted 14 August 2013
Available online 29 August 2013
Keywords:
Eugenia uniora
Pressurised uid extraction
High performance liquid chromatography
Centrifugal partition chromatography
High resolution mass spectrometry

a b s t r a c t
Brazilian cherry seeds are a waste product from juice and frozen pulp production and, the seeds composition was investigated to valorize this by-product. Compounds separation was performed with ethanol
by pressurised uid extraction (PFE). Here we determine the effect of temperature (T), static time (ST),
number of cycles (C), and ush volume (VF) on the yield, composition and total phenolic content (TPC)
of the seed extracts. T, ST and their interaction positively inuenced yield and TPC. Extracts were fractionated by high performance liquid chromatography (HPLC) and centrifugal partition chromatography (CPC).
The collected fractions characterizations were made by electrospray ionisation mass spectrometry
(ESI/MS) and high resolution mass spectrometry (HRMS) indicated the presence of ellagic acid pentoside
and deoxyhexose, quercitrin and kaempferol pentoside. All of these compounds have antioxidant
properties and normally are found in plant extracts. These results conrm that Brazilian cherry seed
extract is a potentially valuable source of antioxidants.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
The Brazilian cherry (Eugenia uniora L.) is a traditional crop in
Southern and Southeastern Brazil, but, in recent years, regular cultivation for commercial purposes has begun in the Northeast. It is a
globular berry, 1.55.0 cm in diameter, with seven to ten longitudinal grooves. During maturation, the exocarp varies from green
to yellow, orange, red or dark red (Bezerra, Silva and Lederman,
2000). The fruit has 69% pulp, 31% seeds and is about 85% water
(Guimares, Holanda, Maia, & Moura F, 1982; Silva, 2006). Among
the tropical fruits, the Brazilian cherry has one of the higher levels
of carotenoids (225.9 lg/g) and vitamin C content of 29.4 mg/
100 g. It also contains large amounts of calcium phosphorus and
anthocyanins. The carotenoids are phytouene, b-carotene, f-carotene, b-cryptoxanthin, c-carotene, lycopene, and rubixanthin.
Lycopene represents 32% of the total carotenoids. Among the phenolic compounds, the dark red Brazilian cherry has 22.50 mg/100 g
of anthocyanins (Bezerra et al., 2000; Filho et al., 2008).
Recently, the volatile compounds present in the Brazilian cherry
leaves, whose biological effects have already been proved, were
found in the fruit as selina-1,3,7(11)-trien-8-one for example (Oliveira, Lopes, Cabral, & Eberlin, 2006). An extract obtained by super Corresponding author. Tel.: +55 19 3565 4268; fax: +55 19 3565 4284.
E-mail address: alelopes@usp.br (A.L. Oliveira).
0308-8146/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2013.08.065

critical uid extraction (SFE) is rich in lycopene and other phenolic

compounds in considerable concentration (Malaman, Moraes,
West, Ferreira, & Oliveira, 2011; Oliveira, Kamimura, & Rabi,
2009). However, for Brazilian cherry seeds, no data in the literature
suggests potential biological activity. A recent study showed that
the seed extract is rich in phenolic compounds with high antioxidant activity, twice as much as some fruits also rich in phenolic
compounds (Bagetti, Facco, Rodrigues, Vizzotto, & Emanuelli,
2009). In the industrial processing of concentrated Brazilian cherry
juice and frozen pulp, the seeds are treated as waste with negligible commercial value, with thousands of tonnes of solid waste produced annually. As numerous biological activities are attributed to
Brazilian cherry leaves and as the characteristic odour of the leaves
is also present in the seeds, it appears important to research the
compounds present in the seeds in order to characterise this
agro-food waste as a source of raw material for food or related
industries.
Another goal of this work was to develop processing systems
that exploit potentially bioactive components. Beyond scientic research, PFE has several advantages for obtaining extracts from natural products (Vandenburg, Clifford, Bartle, Carroll, & Newton,
1999). PFE or accelerated solvent extraction (ASE) gave similar results to Soxhelt in terms of recovery, repeatability and selectivity.
However, both extraction time and solvent consumption were
dramatically reduced with PFE. In fact increased temperature

A.L. Oliveira et al. / Food Chemistry 145 (2014) 522529

decreases the viscosity of the solvent enabling an improved penetration of the matrix and enhances diffusivity of the solvent. Moreover, high pressure keeps the solvent in the liquid state and forces
the solvent through the matrix. The reduced time of extraction
avoids a possible thermal degradation (Kaufmann & Christen,
2002; Wang & Weller, 2006). PFE can be performed using static
or dynamic methods or a combination of both. This method is
widely used as an extraction technique for sample preparation,
to identify the presence of minor components. More recently, this
extraction technique has also been used to obtain extracts enriched
with bioactive compounds (Claude, Morin, Lafosse, Belmont, &
Haupt, 2008; Pl et al., 2007). In this paper, we seeked to optimise
the PFE procedure by examining the impact of altering four important variables: extraction temperature (T), static time (ST), number
of cycles (C) and the volume of solvent in the extractor in each cycle (VF). In addition, chemical analyses were performed to detect
biological activity, as well as analyse composition of these extracts
using thin layer chromatography (TLC), high performance liquid
chromatography (HPLC) coupled with UV, evaporative light scattering detectors (ELSD), and diode-array detection (DAD), centrifugal partition chromatography (CPC) and HPLC followed by mass
spectrometry coupled with electrospray ionisation (ESI/MS).

2. Materials and methods

2.1. Feedstock and chemicals
Native red Brazilian cherries collected from a local farm in Brazil
were selected manually and were separated from the seeds. These
seeds were washed to remove traces of pulp, and dried in forcedair at 50 C for 54 h. Dried seeds were packed in waterproof bags
and stored at 18 C. The ground seeds had 12.73 0.2% water.
The average crushed seeds diameter (0.48 mm) was determined
by shaking for 15 min in six-game series standard Tyler sieves.
For the extraction process anhydrous ethanol was used (Carlo
Erba, Val-de-Reuil, France) as solvent and sodium sulfate (99%
anhydrous; Merck, Val de Fontenay, France) as adsorbent material.
As reagents for preparation of puried extracts methanol and chloroform were used, from Carlo Erba (Val-de-Reuil, France). For spectrophotometric phenols detection the FolinCiocalteu reagent was
used and gallic acid, which purchased from SigmaAldrich (Saint
Quentin Fallavier, France).
For TLC analysis Silica Gel 60 F254 plates were used (Merck,
Darmstadt, Germany) for the stationary phase. Puried water
was from Elgastat UHQ II system (Elga, Antony, France), acetonitrile and ethyl acetate from Carlo Erba (Val-de-Reuil, France); acetic and formic acid were purchased from SigmaAldrich (Saint
Quentin Fallavier, France) and were used in the mobile phase preparation. The anysaldehide acid solution and ninhydrin were from
SigmaAldrich (Saint Quentin Fallavier, France). The Molish reagent was prepared with Naphthol (SigmaAldrich, Saint Quentin
Fallavier, France) and ethanol (Carlo Erba, Val-de-Reuil, France),
the VS1 reagent solution was prepared with H2SO4 and ethanol
both from (Carlo Erba, Val-de-Reuil, France) and Neu-PEG. The
Neu reagent was prepared by mixing diphenyl boric acid ethylamino ester (SigmaAldrich, Saint Quentin Fallavier, France) and
methanol (Carlo Erba, Val-de-Reuil, France), and the PEG reagent
was a solution of polyethylene glycol (PEG) 4000 from Sigma
Aldrich (Saint Quentin Fallavier, France) and ethanol (Carlo Erba,
Val-de-Reuil, France). The developer for determining the antioxidant activity of compounds was composed of 2,2-diphenyl1-(2,4,6 -trinitrophenyl) hydrazyl radical (DPPH) (SigmaAldrich,
Saint Quentin Fallavier, France) solution in methanol (Carlo Erba,
Val-de-Reuil, France). All organic solvents used in this work were
chromatographic grade.

523

2.2. Pressurised uid extraction

PFE was performed using an ASE 100 accelerated solvent
extraction system (Dionex, Voisins le Bretonneaux, France) with
34 mL stainless steel ASE vessels. Anhydrous ethanol was used as
a solvent. Ethanol was chosen for two reasons: to use of a solvent
generally recognised as safe (GRAS), and the Brazilian potential to
produce this solvent. In the optimization process, four parameters
were studied: T, ST, C and VF in a central composite design (CCD)
24, with ve levels, 8 star points and triplicate at the central point
(Table 1). Dried and crushed seeds (5 g) were packed in the vessel
extractor with 5 g sodium sulfate as adsorbent material used to
disperse the vegetal matrix in the extraction cell, allow a better
contact with solvent and, clarify the extract. The ethanol extract
obtained by PFE was named crude extract and was evaporated then
prepared for chromatographic analysis.
2.3. Preparation of crude extracts
The crude extracts were puried to eliminate tannins with a
high degree of polymerization. First, crude extract (100 mg) was
diluted in 2.5 mL methanol and 32.5 mL chloroform. This proportion was important; since preliminary studies showed that any
modication may precipitate other phenolic compounds present
in the extracts (Lhuiller et al., 2007). The diluted chloroform extract
was stored at 4 C for 3 h in the dark. This solution was centrifuged
in a Jouan BR4i multifunction centrifuge (Thermo, Illkrich, France)
at 4000 rpm and 5 C for 10 min. The decanted supernatant was
evaporated under N2 at room temperature in the dark. This solution was named puried ethanol extract.
2.4. Spectrophotometric detection of phenols
The total phenolic content (TPC) in the PFE extracts was determined using the FolinCiocalteu colorimetric method (Singleton &
Rossi, 1965). Crude extracts (1 mL) were diluted in methanol
(1,000 ppm) with 5 mL FolinCiocalteu in water (1:10, v/v). After
10 min, 4 mL anhydrous sodium carbonate solution was added.
After 2 h at room temperature in the dark, the absorbance was
measured on a DU 640 spectrophotometer (Beckman, Villepinte,
France) at wavelengths between 400 and 800 nm. Gallic acid was
used for the calibration curve. The TPC is expressed as the g gallic
acid equivalent (g GAE) per 100 g dry seeds.
2.5. TLC analysis
As the composition of Brazilian cherry seed extract was totally
unknown, TLC was used for preliminary analysis to identify families of compounds in the extracts. We used Silica Gel 60 F254
plates, different mobile phases, developers and standards. The
samples were applied to the plates using a semi-automatic applicator Linomat IV (CAMAG, Muttenz, Switzerland) with 2 lL of each
standard and 5 lL of samples, deposited on the plates in 6 cm
bands separated by 4 cm. Anysaldehide acid solution was used as
developer to analyse the presence of terpenes and sugars in the extract (Wagner & Bladt, 1996). As the presence of sugar was noted in
the extracts, a new TLC method was adopted using three developers. After the chromatographic with acetonitrile:water (75:25, v/v),
ninhydrin was sprinkled on the plate to view amino acid compounds, and then the plate was dried with a hairdryer and warmed
with a heat pistol until the pink bands were visible. Then, Molish
reagent (2 g Naphthol in 100 mL of ethanol) was sprinkled on the
same plate, dried with hairdryer, and then the VS1 reagent solution
(5% H2SO4 in ethanol) was applied. The plate was dried again and
heated with a heat pistol until the violet bands that indicate the
presence of sugars were visible.

524

A.L. Oliveira et al. / Food Chemistry 145 (2014) 522529

Table 1
Matrix of the central composite design (CCD) 24 to study the effect of T, ST, C and VF on the extract yield (Y) and TPC-coded and real variables.

Assay

T (C)

C (no)

VF (%)

ST (min)

Y (%)

TPC (g GAE/100 g)

TPC (ppm)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25a
26a
27a

1 (50)
+1 (70)
1 (50)
+1 (70)
1 (50)
+1 (70)
1 (50)
+1 (70)
1 (50)
+1 (70)
1 (50)
+1 (70)
1 (50)
+1 (70)
1 (50)
+1 (70)
0 (60)
0 (60)
0 (60)
0 (60)
0 (60)
0 (60)
a (40)
+a (80)
0 (60)
0 (60)
0 (60)

1 (2)
1 (2)
+1 (4)
+1 (4)
1 (2)
1 (2)
+1 (4)
+1 (4)
1 (2)
1 (2)
+1 (4)
+1 (4)
1 (2)
1 (2)
+1 (4)
+1 (4)
0 (3)
0 (3)
0 (3)
0 (3)
a (1)
+a (5)
0 (3)
0 (3)
0 (3)
0 (3)
0 (3)

1 (80)
1 (80)
1 (80)
1 (80)
+1 (120)
+1 (120)
+1 (120)
+1 (120)
1 (80)
1 (80)
1 (80)
1 (80)
+1 (120)
+1 (120)
+1 (120)
+1 (120)
0 (100)
0 (100)
a (60)
+a (140)
0 (100)
0 (100)
0 (100)
0 (100)
0 (100)
0 (100)
0 (100)

1 (4)
1 (4)
1 (4)
1 (4)
1 (4)
1 (4)
1 (4)
1 (4)
+1 (8)
+1 (8)
+1 (8)
+1 (8)
+1 (8)
+1 (8)
+1 (8)
+1 (8)
a (2)
+a (10)
0 (6)
0 (6)
0 (6)
0 (6)
0 (6)
0 (6)
0 (6)
0 (6)
0 (6)

6.70
11.20
6.93
7.44
7.01
7.44
6.81
10.94
7.76
11.63
7.06
13.20
6.42
12.00
8.62
12.07
6.06
8.42
7.50
8.50
6.93
13.38
6.78
14.27
9.07
7.90
8.36

1.00
1.33
0.61
0.79
0.99
0.91
0.71
1.21
0.57
1.00
0.72
1.68
0.80
1.60
1.17
1.42
0.42
1.03
0.52
0.80
0.76
1.68
0.98
1.20
1.30
1.16
1.20

53.95
43.33
55.31
48.95
86.96
74.27
65.82
95.13
49.17
56.04
64.05
32.61
37.23
78.27
47.25
40.40
48.23
40.16
48.64
57.63
65.54
74.14
82.39
62.32
84.01
85.03
84.89

Central point of the experimental design (CCD), GAE = gallic acid equivalent, Y% extractg
 100.
seedsg

In order to verify the presence of phenolic compounds in the extracts, another TLC methodology was applied. The mobile phase
was ethyl acetate:acetic acid:formic acid:water (100:11:11:26,
v/v/v/v) with the same stationary phases used to identify sugars,
the plate was developed with Neu-PEG (Wagner & Bladt, 1996).
Neu reagent was obtained by mixing 1 g of diphenyl boric acid ethylamino ester in 100 mL of methanol and PEG reagent was a solution of polyethylene glycol (PEG) 4000 at 5% in ethanol.
Moreover, TLC was used in order to determine the antioxidant
activity of compounds. For this analysis, the developer used was
a solution of 0.05% 2,2-diphenyl-1-(2,4,6-trinitrophenyl) hydrazyl
radical (DPPH) in methanol sprinkled on the plate after elution
and drying. The compounds with anti-oxidant properties reacted
with DPPH radical generating white spots on the plate.

2.6. HPLC analysis

Crude and puried extracts were analysed using a LaChrom
HPLC (Merck Hitachi, Tokyo, Japan) with an L-7100 pump, an L7200 autosampler, an interface (D-7000) for the L-7400 UV-detector set at 254 nm. To determine the maximum absorption of the
components present in the extracts, an Elite HPLC LaChrom
(VWR Hitachi, Tokyo, Japan) consisting of a L-2130 pump, an L2200 autosampler, with a L-2455 diode array detector (DAD) was
used. The UV detection was at 254 and 360 nm and peak spectra
were recorded from 200 to 600 nm.
For the stationary phase, LiChrospher 100 RP-18 (100  4.6 mm
 5 lm; Agilent Technologies, Palo Alto, CA) and LiChrospher C18
(150  4.6 mm  5 lm; Bischoff Chromatography, Leonberg, Germany) columns were used. The mobile phase consisted of ultra
pure water distilled with a Millipore Elix UV system (Millipore,
Saint-Quentin-en-Yvelines, France) and then puried with an Elgastat UHQ II system (Elga, Antony, France) (A) and methanol (B),
both acidied with 0.1% formic acid. The ow (1.0 mL/min) and T
(25 C) were constant. The solvent concentrations varied during

the chromatographic analysis according to the applied gradient

of 90% A for 5 min, decreasing from 90 to 50% A for 15 min, from
50% to 30% A for 10 min, 30 to 0% A for 5 min, and then 0% for
5 min, from 0 to 90% for 5 min, and nally at 90% A for 5 min.

2.7. HPLC/ESI/MS and HRMS identication

The main CPC fractions were analysed by both positive and negative HPLC/ESI/MS using an Agilent 1100 Series HPLC (Agilent)
with an interface for a DAD detector and ESI/MS. HPLC conditions
were used and the mass spectra were obtained on an API3000 triple quadrupole mass spectrometer (AB SCIEX, Foster City, CA, USA)
equipped with a turbo ion spray source and analysis software version 1.4.2 (Applied Biosystem MDS Sciex). The 1 mL/min ow rate
from the HPLC device was split to a ow rate of approximately
0.3 mL/min directed to the MS system. In the analysis of a single
scan (Q1), the quadrupole was operated under the following conditions: ow rate of the curtain gas of 1.2 L/min, the nebulizer gas
1.2 L/min; with +4.2 kV ion spray voltage for positive mode and
4.0 kV for negative mode. The declustering potential (DP) was
70 V, 300 V focusing potential (FP), and 10 V entrance potential
(EP) were used for positive mode, while a DP of 100 V, FP of
400 V and EP of 10 V were used for negative mode. Nitrogen
was used as the curtain and nebulizer gas, and compressed air as
auxiliary gas. The mass spectra were obtained in a scan range from
100 to 1000 m/z. In the analysis of precursor ions and product ion
conditions were as previously described, but the collision energy
(CE) varied over the analysis.
High resolution mass spectrometry (HRMS) was performed on a
maXis mass spectrometer (Bruker, Bremen, Germany). The mass
spectrometer was operated in negative modes and acquired data
in the mass range from m/z 50 to 1650. The capillary voltage was
set at +4.5 kV, the end plate offset at 500 V, the drying N2 ow
at 6 L/min at 200 C, and nebulizer N2 at 1 bar. The accurate mass
data of the molecular ions were processed though the Data

A.L. Oliveira et al. / Food Chemistry 145 (2014) 522529

Analysis 4.0 software (Bruker Daltonik), which provided a list of

possible elemental formulas using the SmartFormula Editor tool.
2.8. Statistical analysis
The inuence of the independent variables studied in the PFE
process using ethanol as the solvent in the extraction yield (Y), in
the TPC, and in the concentration of the main component present
in extracts, was veried using the STATISTICA software version 8
(StatSoft, Tulsa, OK) for Windows.
3. Results and discussion
3.1. PFE
The extraction yield (Y) of the crude PFE extracts and TPC were
the dependent variables analysed as responses in the CCD 24.
Yields were high, considering the raw material, with a mean of
8.90 2.45% (Table 1); consequently, ethanol is a suitable solvent
to obtain large quantities of this material. The evaluation of four
independent variables (T, C, VF and ST) and their interaction
showed T, ST, and their interactions showed signicant effects
(P < 0.05) and positively inuenced the yield. The inuence of T
was higher relatively to ST. Specically, for this material; there
was no inuence of the number of cycles (C) and of the ush volume used in each cycle (VF). Considering the cost and ease of
use, the fact that these variables (C and VF) do not affect the extraction is signicant, because there is no need to use large solvent volumes to generate higher yields. The analysis of variance (ANOVA)
for the 1st and 2nd order experimental designs to analyse the
PFE efciency with ethanol shows that, for both models (linear
and quadratic), there was a signicant effect of these variables
on the yield. The coefcients of determination for linear
(R2 = 0.85) and quadratic (R2 = 0.84) models suggest that both,
when tted to experimental data, have good predictive capacity
for the extract yield. The regression equation to predict Y as a function of independent variables (T, C, VF and ST) are shown in equations 1 and 2 (Eqs. (1) and (2)).

Y 8:88 1:79T  0:89ST 0:59T ST

Y 8:47 1:82T  0:66C  0:79ST 0:59T ST

The regression coefcients, which showed no statistical signicance (P > 0.05), were not considered in the model, but were added
to the pure error. For the linear model (Eq. (1)), T, ST and the

525

interaction between them (ST  T) were signicant variables. The

same is observed in the quadratic model (Eq. (2)). However, the
number of cycles (C) was also signicant. The response surfaces
analysis (RSA) of the linear and quadratic model tted to the experimental values of PFE extract yields for T and ST showed that elevated T raised Y (Fig. 1). The same is observed for ST, as longer
contact time between the solvent and matrix resulted in better
performance.
3.2. TPC
The TPC, expressed as gGAE per 100 g seeds, ranged from 0.42
to 1.68 (mean, 1.02 0.34), showing that the ethanol extracts obtained by PFE had a high concentration of phenolic compounds
compared to the concentration in other tropical fruits seeds (Roesler et al., 2007). All 27 tests (Table 1) presented a similar prole
with maximal absorbance around 750 nm. The concentration of
phenolic compounds in the extracts was a mean of
61.55 17.66 ppm. Statistical analysis also showed that the independent variables T and ST signicantly inuenced the TPC in the
extracts. These variables were also relevant to Y, as the extracts
with higher Y were also rich in phenolic compounds. Unlike for
Y, TPC was also inuenced by C interaction with ST. This is an indication that phenolic compounds strongly attached to seeds need
more contact time and a greater amount of solvent to be extracted
properly. For the TPC, ANOVA showed that, although there was a
signicant effect of T and ST on TPC, the low coefcients of determination calculated suggest that linear (R2 = 0.74) and quadratic
(R2 = 0.71) models tted to experimental data do not have good
predictive ability.
3.3. Extract characterization using TLC
A preliminary TLC analysis identied the compound classes
present in the PFE extracts. The rst compound classes identied
in the extracts were amino acids and carbohydrates (Fig. 2a). Similar ngerprints were observed in ultrasonic methanol extraction
and PFE with methanol, water and formic acid (8:0.5:1.5, v/v/v)
mixture (Bagetti et al., 2009) (samples 3 and 4) while sample 5
highlights other compounds more reactive on silica in the crude
ethanol PFE extract. Some spots have retention close to the phenolic compounds (phloridizin and avicularin used as standards in
samples 6 and 8). Chromatographic plates stained with Neu-PEG
show the presence of phenolic compounds detected at 366 nm
(Fig. 2b). In Fig. 2c, the Neu-PEG reagent enables to show the

Fig. 1. T and ST RSA for the PFE extract yield for the linear and quadratic models (CCD 24).

526

A.L. Oliveira et al. / Food Chemistry 145 (2014) 522529

Fig. 2. Analysis of Brazilian cherry seed extract by TLC. TLC was performed with a Silica Gel 60 F254 plate using Brazilian cherry seed extract as the source material as
described in the Materials and Methods section. (a) Ninhydrin, Molish, and VS1 revelation. The mobile phase was acetonitrile:water (75:25, v/v). 1: Sucrose, 2: Rafnose, 3:
Ext. Ultra sonic (water/methanol/formic acid, 8:0.5:1.5, v/v/v), 4: PFE extracted with water/methanol/formic acid (8:0.5:1.5, v/v/v), 5: PFE extracted with ethanol, 6:
Phloridzin dehydrate, 7: Rutin, 8: Avicularin. (b) Neu-PEG revelation. The mobile phase was ethyl acetate:acetic acid:formic acid:water (100:11:11:26, v/v/v/v), 366 nm
wavelength. 1: Tannins isolated from crude ethanol PFE (5000 ppm), 2: Crude ethanol PFE (5000 ppm), 3: Puried ethanol PFE (10,000 ppm), 5: Puried ethanol PFE
(10,000 ppm), 4, 6 and 7: HPLC F2 of puried ethanol PFE (1000 ppm), (c) and (d): Same mobile phase and samples used in (b) Neu-PEG revelation (track 1, 2, 3, 4 and 5) and
DPPH (0.05%) revelation (track 10 , 20 , 30 , 50 ).

presence of tannins in crude ethanol PFE (samples 1 and 2) and

phenolic compounds in puried ethanol PFE (samples 3 and 5),
and in F2, fraction collected in HPLC from puried ethanol PFE

(sample 4) both in Fig. 2d. The antioxidant activity of the crude

(sample 10 and 20 ) (Fig. 2c) and puried extracts (samples 30 and
50 ) (Fig. 2d) was shown using DPPH (2,2-diphenyl-1-(2,4,6-trinitro-

Fig. 3. Components in Brazilian cherry seed extracts detected by HPLC monitored with a UV detector. (a): crude (blue line) and puried (red line) extracts, LiChrospher RP-18
(100  4.6 mm  5 lm), mobile phase water (A) /methanol (B) both acidied with 0.1% formic acid, at 95% A for 5 min, 9550% A in 5 min, 500% A in 20 min, 0% A for 5 min,
095% A in 5 min, 95% A for 10 min, with UV detection at 254 nm, (b): puried extract from Brazilian cherry seeds, LiChrospher C18 (150  4.6 mm  5 lm), mobile phase
water (A) /methanol (B) both acidied with 0.1% formic acid, at 95% A for 5 min, 9550% A in 5 min, 500% A in 20 min, 0% A for 5 min, UV detection at 360 nm (blue line) and
254 nm (red line). (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)

A.L. Oliveira et al. / Food Chemistry 145 (2014) 522529

527

phenyl) hydrazyl radical) as the developer for TLC. DPPH is a stable

free radical and has been used widely to assess the ability of natural antioxidants to bind free radicals. This qualitative analysis
shows that tannins with a high degree of polymerization present
in the alcohol extract had antioxidant activity (Fig. 2c), as well as
the puried extract (Fig. 2d).

the ion at m/z 255, also typical of quercetin (Sandhu and Gu,
2010). The phenolic compound quercetin 3-O-b-rhamnoside may
be suggested in this fraction. Unfortunately, due to inadequate separation of the peaks by HPLC, mass spectrometry was not effective.
Thus, it was decided to separate the peaks by CPC and then perform HPLC/HRMS analysis.

3.4. Extract characterization using HPLC

3.5. CPC analysis and HPLC/ESI/HRMS

Fig. 3a shows the analysis by HPLC at 254 nm of crude ethanol

PFE (blue line) and Puried ethanol PFE (red line). The two chromatograms indicate a common peak at about 19 min while the
tannins eliminated by purication were found in the dead volume
(blue line). The shape of the common peak eluted at about 19 min
showed the co-elution of two compounds. Therefore, the same
analysis was carried out using a longer column Lichrospher C18
(150  4.6 mm  5 lm) that enables to split the peak of the puried extract in two peaks collected in fractions F1 and F2 (Fig. 3b
shows the chromatogram at 254 and 360 nm). The spectral analysis of F1 by DAD showed two absorbance maxima at 251253 nm
and 360361 nm while the one of F2 was 253255 nm and
366 nm. The asymmetry of the two peaks corresponding at F1
and F2 suggested other co-eluted compounds.
ESI/MS analyses were performed by direct infusion of F1 and F2
in negative mode to identify main compounds of these HPLC fractions. The fragmentation pattern shows a same base peak in F1 and
F2 at m/z 263. In the F2 analysis, the ion at m/z 301 has as a precursor the ion at m/z 447, which could be from the loss of 146 that
suggest a rhamnoside. The ion at m/z 301 is a typical fragment of
quercetin that, with the loss of a fragment at m/z 46, generates

The puried ethanol PFE extract was fractionated by CPC using

the Arizona C system which could provide better separation of
these main compounds (Berthod, Hassoun, & Ruiz-Angel, 2005;
Michel, Destandau, & Elfakir, 2011). The fractions collected in the
tubes 1214 were joined together in the fraction named CPC F3
and the fractions in tubes 4954 in the fraction CPC F6 (Fig. 4).
These two fractions contain the compounds observed in the fractions F1 and F2 obtained previously by HPLC; they were analysed
by HPLC/ESI/HRMS. Using the Smart Formula Editor tool from the
Data Analysis 4.0 software, we identied the principal compounds
presents in these fractions (Table 2). The two peaks observed at
19.3 and 19.5 min in fraction CPC F3 showed a maximum absorbance of 253 and 357 nm, and 253 and 360 nm, respectively. The
rst peak at 19.3 min has been identied as ellagic acid pentoside,
a phenolic acid in conjunction with a sugar (Table 2). This compound, present in plants, has antioxidant properties. In addition
to the tentative identication by HRMS presented here, the literature revealed the presence of this acid in strawberries, with the
same absorbance values and the ion MS/MS at m/z 301 as the principal ion (Aaby, Ekeberg, & Skrede, 2007) the aglycone formed by
loss of a pentoside [MH132] (Liang, Jin, Feng, & Ke, 2011). In

Fig. 4. HPLC chromatogram of the puried ethanol extract of Brazilian cherry seeds (F1 and F2 HPLC fractions) and of its two CPC fractions (CPC F3 tubes 1214, and CPC F6
tubes 4954). Column: LiChrospher C18 (150  4.6 mm  5 lm), with a mobile phase water (A) /methanol (B) both acidied with 0.1% formic acid, at 95% A for 5 min, 9550%
A in 5 min, 500% A in 20 min, 0% A for 5 min, UV detection at 254 nm.

528

A.L. Oliveira et al. / Food Chemistry 145 (2014) 522529

Table 2
HPLC/ESI/HRMS data and tentative identication of the main compounds present in the PFE extracts of Brazilian cherry seeds.
HPLC/CPC

tR (min)

UVvis kmax
(nm)

[M-H]
(amu)

MS/MS

[M] formula

Tentative identication

F1/F3

19.3
19.5

253, 357
253, 360

20.3
20.8
22.2

254, 365
253, 367
263, 316, 353

433.04151
463.08849
447.05704
300.99951
447.09391
417.08456

301
301
301
229
300, 301
285

C19H14O12
C21H20O12
C20H16O12
C14H6O8
C21H20O11
C20H18O10

Ellagic acid pentoside

Quercetin hexoside
Ellagic acid deoxyhexoside
Ellagic acid
Quercitrin
Kaempferol pentoside

F2/F6

the second peak at 19.5 min, there were two compounds, a derivative of quercetin hexose and an ellagic acid deoxyhexose with the
same retention time and both with the same maximum absorbance (253 and 360 nm). The quercetin hexoside is a avonol glycoside, also an antioxidant compound present in plants (Lin, Chen,
& Harnly, 2008; Mtt, Kamal-Eldin, & Trrnen, 2003). The product ion at m/z 301 is the quercetin aglycone formed by loss of a
hexoside [MH162] (Yu et al., 2008). Ellagic acid deoxyhexoside, another ellagic acid derivative with the parent ion at m/z
447 that was identied as an antioxidant compound in chestnut
(Sanz et al., 2010). For this compound, the product ion at m/z
301 is the ellagic acid aglycone formed by loss of a deoxyhexoside
[MH146]. Even if preliminary analysis of HPLC F1 by HPLC/ESI/
MS showed more than one component in the single chromatographic peak, the most intense ion spectra at m/z 301 can correspond to the quercetin or to ellagic acid structure. So, this
analysis was insufcient to propose a structure for the different
compounds. The detection of these several compounds in the same
peak was only possible thanks to the sensitivity and precision of
the HRMS which allow to detect several [MH] ions at different
m/z that all lead after fragmentation to the m/z 301 product ion.
According to its exact molecular mass obtained by HRMS, the
corresponding molecular formula peak at 20.3 min could be identied as ellagic acid.
In CPC F6 fraction, two compounds were detected at 20.8 and
22.2 min, quercitrin and a kaempferol pentoside (Table 2). Both
compounds have antioxidant properties and are known constituents of plant extracts. In the HPLC/ESI/MS analysis of HPLC F2,
the ion at m/z 301 has a precursor ion [M-H] at m/z 447, indicating that it was derived from the loss of a rhamnoside (a deoxyhexose) [M-H-146]. It could be that the phenolic compound quercetin
3-O-b-rhamnoside or quercitrin with [M-H] ion at m/z 447 is
present in the Brazilian cherry seed PFE extract, which was conrmed by the HRMS analysis (Table 2) and by HPLC/MS analysis
of quercitrin standard in the same conditions.
The avonol kaempferol, also present in plants, has antioxidant
activity. Like the others, this avonol glycoside is also present in
other seeds, such as in green beans (Price, Colquhoun, Barnes, &
Rodhes, 1998) and lotus seeds (Husam et al., 2010). Its presence
in Brazilian cherry seed extracts was identied tentatively by UV
spectrum (263 and 353 nm) and HRMS analysis. The abundant
product ion at m/z 285 is the kaempferol aglycone that lost a
pentoside [M-H-132] (Table 2).
The tentative identication of the composition of Brazilian cherry seed extracts, obtained by PFE with ethanol as solvent, indicated
the presence of avonol and acids conjugated with sugars. These
components have antioxidant properties, which were conrmed
by the high antioxidant activity of these extracts. Furthermore,
the tannins with a high degree of polymerization, removed from
the crude extract, may also have the same action. Experiments
are now being conducted to examine their possible bio-applications. Also in this study, PFE was optimised at laboratory scale; it
will be interesting to demonstrate that this process, can be used
at large scale to obtain extracts rich in bioactive compounds and

to generate products from agro-processing waste, abundant in Brazil. Moreover, ethanol is a GRAS solvent and has easy accessibility
in Brazil which is one of the largest producers.
Acknowledgement
The authors would like to thank FAPESP (State of So Paulo Research Foundation, Brazil) Project 2008/00148-0 for nancial
support.
References
Aaby, K., Ekeberg, D., & Skrede, G. (2007). Characterization of phenolic compounds
in strawberry (fragaria x ananassa) fruits by different HPLC detectors and
contribution of individual compounds to total antioxidant capacity. Journal of
Agricultural and Food Chemistry, 55(11), 43954406.
Bagetti, M., Facco, E. M. P., Rodrigues, D. B., Vizzotto, M., & Emanuelli, T. (2009).
Antioxidant capacity and composition of pitanga seeds. Cincia Rural, 39(8),
25042510.
Berthod, A., Hassoun, M., & Ruiz-Angel, M. (2005). Alkane effect in the Arizona
liquid systems used in countercurrent chromatography. Analytical &
Bioanalytical Chemistry, 383, 327340.
Bezerra, J. E. F, Silva, J. F, Jr, & Lederman, I. E. (2000). Pitanga (Eugenia uniora L.) Srie
Frutas Nativas. Jaboticabal, So Paulo, BR: Funep.
Claude, B., Morin, P., Lafosse, M., Belmont, A.-S., & Haupt, K. (2008). Selective solidphase extraction of a triterpene acid from a plant extract by molecularly
imprinted polymer. Talanta, 75(2), 344350.
Filho, G. L., De Rosso, V. V., Meireles, M. A. A., Rosa, P. T. V., Mercadante, A. Z.,
Oliveira, A. L., et al. (2008). Supercritical CO2 extraction of carotenoids from
pitanga fruits (Eugenia uniora L.). The Journal of Supercritical Fluids, 46(1),
3339.
Guimares, F. A., Holanda, L. F. F., Maia, G. A., & Moura F, J. A. (1982). Estudos
analticos e fsicos em polpa e semente de pitanga (Eugenia uniora L.). Cincia e
Tecnologia de Alimentos, 2, 208215.
Husam, M. K., Huang, D., Xie, B., He, H., Yang, E., Tian, B., et al. (2010). Flavonols of
lotus (Nelumbo nucifera, Gaertn.) seed epicarp and their antioxidant potential.
European Food Research & Technology, 231(3), 387394.
Kaufmann, B., & Christen, P. (2002). Recent extraction techniques for natural
products: Microwave-assisted extraction and pressurised solvent extraction.
Phytochemical Analysis, 13, 105113.
Lhuiller, A., Fabre, N., Moyano, F., Martins, N., Claparols, C., Fourast, I., et al. (2007).
Comparison of avonoid proles of Agauria salicifolia (Ericaceae) by liquid
chromatography-UV diode array detectionelectrospray ionisation mass
spectrometry. Journal of Chromatography A, 1160(12), 1320.
Liang, X.-m, Jin, Y., Feng, J.-t, & Ke, Y.-x. (2011). Identication and structure
elucidation of compounds from herbal medicines. In W. J. H. Liu (Ed.).
Traditional Herbal Medicine Research Methods: Identication, Analysis, Bioassay,
and Pharmaceutical and Clinical Studies (pp. 448). Hoboken, NJ, USA: John Wiley
& Sons.
Lin, L.-Z., Chen, P., & Harnly, J. M. (2008). New phenolic components and
chromatographic proles of green and fermented teas. Journal of Agricultural
and Food Chemistry, 56(17), 81308140.
Mtt, K. R., Kamal-Eldin, A., & Trrnen, R. (2003). High-performance liquid
chromatography (HPLC) analysis of phenolic compounds in berries with diode
array and electrospray ionization mass spectrometric (MS) Detection: Ribes
Species. Journal of Agricultural and Food Chemistry, 51(23), 67366744.
Malaman, F. S., Moraes, L. B., West, C., Ferreira, N. J., & Oliveira, A. L. (2011).
Supercritical uid extracts from the Brazilian cherry (Eugenia uniora L.):
Relationship between the extracted compounds and the characteristic avour
intensity of the fruit. Food Chemistry, 124(1), 8592.
Michel, T., Destandau, E., & Elfakir, C. (2011). On-line hyphenation of centrifugal
partition chromatography and high pressure liquid chromatography for the
fractionation of avonoids from Hippopha rhamnoides L. berries. Journal of
Chromatography A, 1218(9), 61736178.
Oliveira, A. L., Kamimura, E. S., & Rabi, J. A. (2009). Response surface analysis of
extract yield and avour intensity of Brazilian cherry (Eugenia uniora L.)

A.L. Oliveira et al. / Food Chemistry 145 (2014) 522529

obtained by supercritical carbon dioxide extraction. Innovative Food Science &
Emerging Technology, 10(2), 189194.
Oliveira, A. L., Lopes, R. B., Cabral, F. A., & Eberlin, M. N. (2006). Volatile compounds
from pitanga fruit (Eugenia uniora L.). Food Chemistry, 99(1), 15.
Pl, J., Ostr, E. V., Karsek, P., Roth, M., Benesov, K., Kotlarikova, P., et al. (2007).
Comparison of two different solvents employed for pressurised uid extraction
of steviosde from Stevia rebaudiana: methanol versus water. Analytical &
Bioanalytical Chemistry, 388(8), 18471857.
Price, K. R., Colquhoun, I. J., Barnes, K. A., & Rodhes, M. J. C. (1998). Composition and
content of avonol glycosides in green beans and their fate during processing.
Journal of Agricultural and Food Chemistry, 46(12), 48984903.
Roesler, R., Malta, L. G., Carrasco, L. C., Holanda, R. B., Sousa, C. A., & Pastore, G. M.
(2007). Atividade antioxidante de frutas do cerrado. Cincia e Tecnologia de
Alimentos, 27(1), 5360.
Sandhu, A. K., & Gu, L. (2010). Antioxidant capacity, phenolic content, and proling
of phenolic compounds in the seeds, skin, and pulp of Vitis rotundiolia
(Muscadine grapes) as determined by HPLC-DAD-ESI-MS. Journal of Agricultural
and Food Chemistry, 58(8), 46814692.
Sanz, M., Cadaha, E., Esteruelas, E., Muos, A. M., Simn, B. F., Hernanndez, T., et al.
(2010). Phenolic Compounds in Chestnut (Castanea sativa Mill.) Heartwood.

529

Effect of Toasting at Cooperage. Journal of Agricultural and Food Chemistry,

58(17), 96319640.
Silva, S. M. (2006). Pitanga. Revista Brasileira de Fruticulture, 28(1), 1820.
Singleton, V. L., & Rossi, J. A. Jr, (1965). Colorimetry of total phenolics with
phosphomolybdicphosphotungstic acid reagents. American Journal of Enology
and Viticulture, 16(3), 144158.
Vandenburg, H. J., Clifford, A. A., Bartle, K. D., Carroll, J., & Newton, I. D. (1999).
Comparation of pressurised uid extraction and microwave assisted extraction
with atmospheric pressure methods for extraction os additives from
polypropylene. Analyst, 124(3), 397400.
Wagner, H., & Bladt, S. (1996). Plant Drug Analysis A Thin Layer Chromatography
Atlas. London, UK: Springer.
Wang, L., & Weller, C. L. (2006). Recent advances in extraction of nutraceuticals
from plants. Trends in Food Science & Technology, 17, 300312.
Yu, J., Yuan-sheng, X., Fei-fang, Z., Xue, X. -y., Xu, Q., & Liang, X. -m. (2008).
Systematic screening and characterization of avonoid glycosides in Carthamus
tinctorius L. by liquid chromatography/UV diode-array detection/electrospray
ionization tandem mass spectrometry. Journal of Pharmaceutical and Biomedical
Analysis, 46(3), 418430.