Food Chemistry 197 (2016) 266–272

Contents lists available at ScienceDirect

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

Conventional, ultrasound-assisted, and accelerated-solvent extractions

of anthocyanins from purple sweet potatoes
Zhan Cai a,1, Ziqian Qu a,1, Yu Lan b, Shujuan Zhao a, Xiaohua Ma b, Qiang Wan a, Pu Jing a,⇑, Pingfan Li c,⇑
a
Research Center for Food Safety and Nutrition, Key Lab of Urban Agriculture (South), Bor S. Luh Food Safety Research Center, School of Agriculture & Biology, Shanghai Jiao
Tong University, Shanghai 200240, China
b
Hainan Yedao Group Co. Ltd., Hainan 570105, China
c
Guangdong Industry Technical College, Guangdong 510300, China

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

Article history: Purple sweet potatoes (PSPs) are rich in anthocyanins. In this study, we investigated the extraction
Received 22 July 2015 efficiency of anthocyanins from PSPs using conventional extraction (CE), ultrasound-assisted extraction
Received in revised form 19 October 2015 (UAE), and accelerated-solvent extraction (ASE). Additionally, the effects of these extraction methods
Accepted 24 October 2015
on antioxidant activity and anthocyanin composition of PSP extracts were evaluated. In order of
Available online 10 November 2015
decreasing extraction efficiency, the extraction methods were ASE > UAE > CE for anthocyanins
(218–244 mg/100 g DW) and CE > UAE > ASE for total phenolics (631–955 mg/100 g DW) and
Chemical compounds studied in this article:
flavonoids (28–40 mg/100 g DW). Antioxidant activities of PSP extracts were CE
UAE > ASE for ORAC
Cyanidin 3-sophoroside-5-glucoside
(PubChem CID: 44256732)
(766–1091 mg TE/100 g DW) and ASE > CE
UAE for FRAP (1299–1705 mg TE/100 g DW). Twelve
Peonidin 3-sophoroside-5-glucoside anthocyanins were identified. ASE extracts contained more diacyl anthocyanins and less nonacyl and
(PubChem CID: 44256845) monoacyl anthocyanins than CE and ASE extracts (P < 0.05).
Ó 2015 Elsevier Ltd. All rights reserved.
Keywords:
Extraction efficiency
Anthocyanins
Phenolics
Flavonoids
Antioxidant activity

1. Introduction Rodriguez-Solana, Salgado, Dominguez, & Cortes-Dieguez, 2015).

Anthocyanins are flavonoids responsible for the red, purple,
violet, and blue colors of fruits, vegetables, and cereals (Mazza &
Miniati, 1993). The increasing interest in anthocyanins as food
colorants is attributed to the safety concerns of synthetic food
color additives (Giusti & Jing, 2007; Wrolstad & Culver, 2012).
Purple sweet potatoes (Ipomoea batatas L.), which are rich in
anthocyanins, have been grown in China as food and food colorants
(Fan, Han, Gu, & Chen, 2008). Anthocyanins from purple sweet
potatoes (PSPs) have bioactive effects, including antioxidant (Teow
et al., 2007) and anti-inflammatory activities (Zhang et al., 2009).
Additionally, PSP anthocyanins attenuate dimethylnitrosamine-
induced liver injury in rats (Hwang et al., 2011).
Extraction is a separation technique that affects the yield, qual-
ity, and composition of target compounds (Jing & Giusti, 2007;
⇑ Corresponding authors.
E-mail address: pjing@sjtu.edu.cn (P. Jing).
1
These authors made equivalent contributions to the manuscript.
An effective extraction method maximizes the extraction of
target compounds with minimal degradation and is based on
environmental friendly technologies.
The conventional extraction (CE) of anthocyanins, which is
based on the bath stirring extraction method, is performed in
acidic solutions to obtain the red stable flavylium cation (Santos,
Veggi, & Meireles, 2010). Compared to CE, ultrasound-assisted
extraction (UAE), which has been used in the extraction of antho-
cyanins (Lien, Chan, Lai, Huang, & Liao, 2012) and other phenolics
(Carrera, Ruiz-Rodriguez, Palma, & Barroso, 2012), have several
advantages including reduced processing time and solvent volume.
Additionally, UAE has an effective mass transfer and solvent
penetration through the disruption of plant cell walls via acoustical
cavitation (Rastogi, 2011). Accelerated-solvent extraction (ASE) is a
solvent extraction technique that is performed at elevated temper-
atures and pressure, conditions that improve extraction kinetics
and reduce extraction time and solvent volume (Abdel-Aal,
Akhtar, Rabalski, & Bryan, 2014; Truong, Hu, Thompson, Yencho,
& Pecota, 2012).

http://dx.doi.org/10.1016/j.foodchem.2015.10.110
0308-8146/Ó 2015 Elsevier Ltd. All rights reserved.
 Z. Cai et al. / Food Chemistry 197 (2016) 266–272 267

In this study, we investigated the extraction efficiency of antho- Table 1

cyanins from PSPs using CE, UAE, and ASE. The effects of these Factors and levels of Taguchi orthogonal array designs.

methods on anthocyanin composition and antioxidant activity Design Methods Factor (i) Level (j)
were evaluated. 1 2 3
L9 (34) CE Temperature (°C) 60 70 80
Time (min) 90 120 150
2. Materials and methods Ethanol (%, v/v) 70 80 90
HCl (%, v/v) 0.01 0.05 0.1
2.1. Materials and reagents UAE Temperature (°C) 40 50 60
Time (min) 45 60 75
Ethanol (%, v/v) 80 90 100
Purple sweet potatoes (PSPs) were provided by Hainan Yedao
Power (W) 200 240 280
Ltd. (Hainan, China). Pure standards of gallic acid, quercetin, 6-hy
L4 (23) ASE Temperature (°C) 80 90
droxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (Trolox),
Static time (min) 15 20
fluorescein sodium salt, Folin–Ciocalteu reagent (FC reagent), Static cycle 1 2
2,20 -azobis dihydrochloride (AAPH) and 2,4,6-Tripyridyl-s-triazine
(TPTZ) were purchased from Sigma–Aldrich (Shanghai, China). All
other chemicals and solvents were of the highest commercial grade 2.3. Total monomeric anthocyanins
and were purchased from Anpel (Shanghai, China).
The total monomeric anthocyanins in PSPs extracts using differ-
2.2. Anthocyanins extraction ent extracted methods were determined by the pH differential
method (Giusti & Wrolstad, 2001). An L5S UV–visible spectropho-
2.2.1. Conventional solvent extraction tometer (Shanghai Analytical Instrument, China) was used to read
PSPs were cut into pieces after wash and prepared in a JJ-2 absorbance at 520 and 700 nm. Monomeric anthocyanins content
blender (Jintan, China) into slurry. Ten grams of PSPs slurry were were expressed as cyanidin-3-glucoside equivalents (CGE), using
placed into a 250 mL flask containing 100 mL aqueous ethanol a molecular weight of 449.2 mg/L and the molar absorptivity of
(70%, 80%, and 90%, v/v). The orthogonal array design L9 (34) was 26,900 L cm1 mg1. Cuvettes with 1-cm path length were used.
employed to study effects of temperature (60, 70 and 80 °C), Measurements were performed in triplicates.
extraction time (90, 120, and 150 min), aqueous ethanol (70%,
80%, and 90%, v/v), and HCl concentration (0.1%, v/v) in Table 1. 2.4. Total phenolics
Then, mixtures were centrifuged in L535R (Xiangyi, China) at
4000 rpm at 4 °C for 15 min and the supernatants were collected. Total phenolic content was measured using a modified
The solution was evaporated using a RE-52 rotary evaporator Folin–Ciocalteu method (Waterhouse, 2001). Briefly, a series of
(Yarong, China) and adjusted to 25 mL with corresponding extrac- tubes were prepared with 0.5 mL water or samples and 2.5 mL
tion solutions. The extracts were stored at 18 °C in the dark until Folin–Ciocalteu reagent for 5 min. After that, 2 mL of 75 g/L Na2CO3
analysis. All experiments were conducted in triplicates. solution was added to each test tube and mixed well before
incubation at room temperature for 2 h in dark. Then the
2.2.2. Ultrasound-assisted extraction absorbance of samples and standards was measured at 760 nm in
Ten grams of PSPs slurry were added into a 250-mL conical the L5S UV–visible spectrophotometer after zeroing the
bottles containing 100 mL of 80%, 90%, or 100% (v/v) aqueous spectrophotometer with a water blank. Total phenolic content
ethanol solutions. The bottles were then capped and put into a was calculated as gallic acid equivalents (GAE) based on a gallic acid
THC ultrasound extraction (Ji Ning Tianhua Ultrasonic Electronics, standard curve. Disposable cuvettes of 1-cm path length were used.
China). Four three-level factors including temperatures (40, 50, and Each sample was evaluated using three replications.
60 °C), duration (40, 50, and 60 min), ethanol concentration (80%,
90%, and 100%, v/v) and ultrasonic output power (200, 240, and 2.5. Total flavonoids
280 W) were studied via the Taguchi orthogonal array design L9
(34) in Table 1. Aqueous ethanol solution was acidified by adding The total flavonoids content of each PSP extracts was deter-
0.1% (v/v) HCl for all UAE experiments. Mixtures were centrifuged mined using a slightly modified method described previously
at 4000 rpm at 4 °C for 15 min after extraction. The supernatants (Meda, Lamien, Romito, Millogo, & Nacoulma, 2005). Briefly, 5 mL
were collected and adjusted to 25 mL after evaporation. All exper- of 2% aluminum trichloride (AlCl3) in methanol was mixed with
iments were conducted in triplicates. The extracts were stored at the same volume of a sample solution and let stand for 10 min.
18 °C in the dark until analysis. Then, the absorbance of samples or standards was measured at
415 nm in the L5S UV–visible spectrophotometer against a blank
2.2.3. Accelerated-solvent extraction sample consisting of a 5 mL methanol solution without AlCl3. The
The accelerated-solvent extractions was carried out with an ASE total flavonoids content was calculated using a standard curve
350 Accelerated Solvent Extractor system (Dionex, Sunnyvale, with quercetin (0–50 mg/L) as the standard. The mean of three
USA). Ten grams of PSPs slurry were added into a 100-mL Zirco- readings was used and expressed as mg of quercetin equivalents
nium extraction cell (Dionex, Sunnyvale, USA). The extraction cells (QE) per 100 g of PSPs.
were arranged in the sample carousel and extracted in a Taguchi L4
(23) design including three two-level factors in Table 1. Factors and 2.6. Antioxidant activity assays
levels have been chosen as a result of considering the cost of
conducting experiments and single factorial experimental results. 2.6.1. Oxygen radical absorbance capacity (ORAC) assay
The 80% (v/v) aqueous ethanol containing 0.1% (v/v) HCl was used The oxygen radical absorbance capacity of the extracts from
as the extraction solvent for all experiments. The supernatants PSPs was investigated according to previously reported procedure
were collected and adjusted to 25 mL after evaporation. Three (Jing et al., 2014) with a slight modification in an Infinite F200 Pro
trails were conducted for all experiments. The extracts were stored microplate reader (Tecan, Männedorf, Switzerland). Samples and
at 18 °C in the dark until analysis. Trolox standards were prepared with ethanol. All other reagents
268 Z. Cai et al. / Food Chemistry 197 (2016) 266–272

were prepared in 75 mmol/L phosphate buffer (pH 7.4). Briefly,
each well in 96-well plate contained 25 lL sample or ethanol for
blank and 150 lL, 4  106 mmol/L fluorescein solution. The plate
with cover was incubated for 30 min in 37 °C, and then 25 lL,
0.153 mol/L AAPH were added to each well to start reaction. The
fluorescence was recorded every minute for 120 min (excitation/
emission: 485/535 nm) at 37 °C. Trolox equivalents were calcu-
lated using the relative area under the curve for samples compared
to a Trolox standard curve prepared under the same experimental
conditions. Results were expressed as Trolox equivalents (TE) per
100 g of PSPs. All experiments were performed in triplicates.
2.6.2. Ferric reducing antioxidant capacity (FRAP) assay
The FRAP method was carried out based on the reduction of
Fe3+-TPTZ to a blue colored Fe2+-TPTZ with modification (Jing
et al., 2012). Briefly, a portion of 10 mmol/L TPTZ was mixed with
the same volume of 20 mmol/L FeCl3  6H2O, and 10 times higher
volume of 300 mmol/L acetate buffer (pH 3.6). The mixture was
incubated at 37 °C for 30 min. Then, 3 mL of FRAP reagent,
100 lL of sample or standards and 300 lL of distilled water were
added to the test tubes and incubated. The absorbance was
measured at 593 nm in the L5S UV–visible spectrophotometer.
Trolox was used as standard for comparison and adequate dilution
of sample was performed. The results were reported as Trolox
equivalents (TE) per 100 g of PSPs. All experiments were conducted
for three times.
2.7. Qualitative and quantitative analysis of PSP anthocyanin-rich
extracts
Anthocyanin-rich extracts (1 mL) were semi-purified using the
method of our previous study (Jing et al., 2012). About 1 mL of
anthocyanins extract was loaded onto a C18 Sep-Pak solid cartridge
(ANPEL, Shanghai, China), which was preconditioned with 2
column volumes of methanol and 3 column volumes of 0.01%
HCl–water (v/v). Sugars and other polar compounds were removed
with 3-column volumes of 0.01% HCl–water, and anthocyanins
and other phenolics were bound to the C18 cartridge. Finally,
anthocyanins were recovered from the cartridge with three
column volumes of 0.01% HCl–methanol. The methanol was
removed by rotary evaporation at 40 °C and the residue was taken
up to about 1 ml with deionized water. The samples were stored at
18 °C for LC–MS and HPLC analysis.
The determination of high-resolution masses of anthocyanins in
extracts was carried out in a Waters Acquity UPLC system (Waters
Corporation, Milford, MA, USA) coupled to the Waters Synapt
HDMS Q-TOF MS detector. Five microliter injected on a COSMOSIL
C18-MS-II column (150  3.0 mm, 5 lm), and the mobile phase
consisted of solvents A (formic acid: water = 1:100, v/v) and B
(formic acid: acetonitrile = 1:100, v/v), using the following
gradient: 5% B between 0 and 0.5 min, 5–20% B between 0.5 and
6 min, 20–35% B between 6 and 8.5 min, 35–50% B between 8.5
and 9.5 min, 50–100% B between 9.5 and 11 min. Peak detection
was carried out online by electrospray ionization in the positive
mode. The applied electrospray/ion optics parameters were
set as follows: capillary voltage, 3.0 kV; sampling cone, 35 V;
collision energy, 4 eV; source temperature, 100 °C; desolvation
temperature, 300 °C; desolvation gas, 500 L/h. Spectra were
collected using full ion scan mode over the mass-to-charge (m/z)
range 200–2000 au. Scan time, 0.3 s; inter scan time, 0.02 s.
Quantitative analysis of anthocyanins were conducted using the
Finnigan Surveyor Plus system (Thermo Scientific, USA). Separation
was achieved by reverse phase elution on a Shim-pack VP-ODS
column (4.6 mm  250 mm, 5 lm, Shimadzu, Kyoto, Japan). The
chromatographic conditions were: flow rate 1 mL/min, sample
injection volume of 10 lL and mobile phase A (formic acid/water,
1:100, v/v) and mobile phase B (formic acid/acetonitrile, 1:100,
v/v). A gradient program was used as follows: 0–5 min, 15% B;
5–30 min, 15–20% B; 30–45 min, 20–40% B; 45–50 min, 40% B.
Spectral information over the wavelength range of 200–800 nm
was collected. And the detection wavelength was set at 530 nm.
Each extract was analyzed in triplicates.
2.8. Experimental design and statistical analysis
The Taguchi orthogonal array design was applied to determine
effects of extraction variables and their main effects. Four three-
level or three two-level factors were selected based on results of

Table 2
Total anthocyanin content of purple sweet potatoes using conventional extraction method (CE) at various levels of temperature, time, ethanol concentration and HCl ratio in
solvent.

Run # Temperature (°C) Time (min) Ethanol (%, v/v) HCl concentration (%, v/v) Monomeric anthocyaninsa (mg/100 g DW)
1 60 90 70 0.01 200.92 ± 15.09 cd
2 60 120 80 0.05 208.04 ± 7.24 cd
3 60 150 90 0.1 204.06 ± 6.40 cd
4 70 90 80 0.1 215.29 ± 6.88d
5 70 120 90 0.01 140.06 ± 9.78a
6 70 150 70 0.05 162.52 ± 1.81ab
7 80 90 90 0.05 179.79 ± 16.54bc
8 80 120 70 0.1 155.88 ± 5.55ab
9 80 150 80 0.01 165.66 ± 14.49ab
ANOVA analysis
SS 7267.250 4443.882 3036.794 2411.389
Df 2 2 2 2
F 33.313 20.371 13.921 11.054
P value 0.000⁄⁄ 0.000⁄⁄ 0.000⁄⁄ 0.001⁄⁄
Range analysis
Ki1 204.34 198.67 173.11 168.88
Ki2 172.62 167.99 196.33 183.45
Ki3 167.11 177.41 174.64 191.74
Rj 37.23 30.67 23.22 22.86

Kij shows the average response of each factor at different levels where i represent a factor and j a level. Rj = Ki.max–Ki.min for each factor.
SS: sum of squares.
Df: degree of freedom.
a
Means within a column for amount of monomeric anthocyanins followed by the same letter are not significantly different at P < 0.05 (Tukey HSD test).
 Z. Cai et al. / Food Chemistry 197 (2016) 266–272 269

their single factorial assays and listed as Taguchi L9 (34) or L4 (23) Table 4
orthogonal design in Table 1, respectively. Three trails have been Yields of monomeric anthocyanins from PSPs extracts using accelerated-solvent
extraction (ASE) at two-level temperature, static time and static cycle.
conducted for each experiment. Statistical analysis was performed
with an analysis of variance (ANOVA) and a range analysis. Run # Temperature Static time Static Monomeric anthocyaninsa
The univariate ANOVA in General Linear Model was applied to (°C) (min) cycle (mg/100 g DW)

determine main effects using SPSS (version16.0, SPSS Inc., Chicago, 1 80 20 2 185.48 ± 2.99a
IL, USA). The Tukey HSD test was used to identify differences in 2 80 15 1 142.46 ± 1.83a
3 90 20 1 197.14 ± 8.33b
means. Tests were conducted in triplicate determinations with 4 90 15 2 252.34 ± 10.59c
data reported as mean ± standard deviation.
ANOVA analysis
Range analysis was used to confirm the effect of each factor and SS 11078.979 111.264 7235.376
determine the optimal level of different factors. The range analysis Df 1 1 1
was applied to parameter optimization. The average response for F 228.638 2.296 149.317
each factor was computed at each level, and labeled as Kij, where P value 0.000⁄⁄ 0.168 0.000⁄⁄
i represent a factor and j a level. The Rj represented the difference Range analysis
of Ki.max and Ki.min was calculated, where Ki.max and Ki.min were the Ki1 163.97 198.90 169.80
Ki2 224.74 191.31 218.91
largest and smallest values among the Kij for each factor, respec-
Rj 60.77 7.59 49.11
tively. According to the largest donating rule (Rao, Kumar,
Prakasham, & Hobbs, 2008), the largest Rj value indicates the most Kij shows the average response of each factor at different levels where i represent a
factor and j a level. Rj = Ki.max–Ki.min for each factor.
significant influence to the monomeric anthocyanins yields. The
SS: sum of squares.
optimal operation parameters were determined when the highest Df: degree of freedom.
monomeric anthocyanin yield among Ki1, Ki2, and Ki3 for each fac- a
Means within a column for amount of monomeric anthocyanins followed by the
tor was clearly distinguished. same letter are not significantly different at P < 0.05 (Tukey HSD test).

Anthocyanin yield from CE ranged from 140.06 to

3. Results and discussion
215.29 mg/100 g DW (Table 2). In CE, the extraction parameter con-
tributions on anthocyanin yield were temperature > time > ethanol
3.1. Anthocyanin extraction efficiency from PSP
concentration > HCl concentration (P < 0.001). Additionally, range
analysis was applied for extraction parameter optimization. Based
This study evaluated anthocyanin extraction efficiency from
on the Ri values, the significance of the extraction parameters in
PSPs using CE, UAE, and ASE. In CE, aqueous ethanol was used as
the range analysis was consistent with the ANOVA results;
a solvent to assess the effects of temperature, time, and HCl con-
therefore, temperature was the most significant factor, while HCl
centration on the anthocyanin extraction efficiency from PSPs.
concentration was the least significant one. The theoretical
Tables 2–4 show the Taguchi’s orthogonal array results, which
maximum anthocyanin yield at 60 °C, 90-min extraction time,
were subjected to ANOVA to determine the significance of main
80% (v/v) ethanol, and 0.1% HCl (v/v) had the highest Kij values
effects, followed by range analysis for optimizing the extraction
among all factor levels.
parameters of each method.
Anthocyanin yield from UAE ranged from 171.94 to
214.92 mg/100 g DW (Table 3). Based on the ANOVA results,
Table 3
Yields of monomeric anthocyanins from PSPs extracts using ultrasound-assisted
temperature, ethanol concentration, and power were significant
extraction method (UAE) at various levels of temperature, time, ethanol concentration
and power.

Run # Temperature Time Ethanol Power Monomeric Table 5

(°C) (min) (%, v/v) (w) anthocyaninsa Total amounts of anthocyanins, phenolics, and flavonoids, and antioxidant activities
(mg/100 g DW) using CE, UAE, and ASE at optimized conditions.
1 40 45 80 200 183.53 ± 10.99ab
Extraction methods
2 50 60 80 240 198.26 ± 10.62bcd
3 60 75 80 280 182.56 ± 4.59ab CE UAE ASE
4 60 60 90 200 214.92 ± 11.59d
Optimized parameters
5 40 75 90 240 190.53 ± 1.33abc
Temperature (°C) 60 50 90
6 50 45 90 280 204.30 ± 11.83bcd
Time/static time 90 45 15
7 50 75 100 200 209.37 ± 13.04 cd
(min)
8 60 45 100 240 205.63 ± 1.33bcd
Ethanol (%, v/v) 80 90 80
9 40 60 100 280 171.94 ± 1.57a
HCl (%, v/v) 0.1 0.1 0.1
ANOVA analysis Power (W) n.a. 200 n.a.
SS 2562.036 65.877 1030.588 1283.768 Static cycle n.a. n.a. 2
Df 2 2 2 2
Extractive efficiency
F 16.354 0.421 6.579 8.195
Monomeric 217.58 ± 2.90a 229.41 ± 4.59b 244.07 ± 11.84c
P value 0.000⁄⁄ 0.663 0.007⁄⁄ 0.003⁄⁄
anthocyaninsa
Range analysis Total phenolics b 955.05 ± 20.90a 769.65 ± 1.50b 630.67 ± 6.54c
Ki1 182.00 197.82 188.12 202.61 Total flavonoids c 40.54 ± 0.35a 32.90 ± 0.56b 27.81 ± 0.96c
Ki2 203.98 195.04 203.25 198.14 ORAC d 1091.06 ± 60.24a 1036.94 ± 110.80a 766.13 ± 114.24b
Ki3 201.04 194.15 195.65 186.27 FRAP d 1299.43 ± 2.32a 1303.14 ± 1.22a 1704.76 ± 6.38b
Rj 21.98 3.67 15.13 16.34
Values are represented as means ± SD (n = 3); within each row, means with the
Kij shows the average response of each factor at different levels where i represent a different letter are significantly different (P 6 0.05).
a
factor and j a level. Rj = Ki.max–Ki.min for each factor. Total monomeric anthocyanins was calculated as cyanidin-3-glucoside equiv-
SS: sum of squares. alents (mg CGE/100 g DW).
b
Df: degree of freedom. Total phenolics were calculated as gallic acid equivalents (mg GAE/100 g DW).
a c
Means within a column for amount of monomeric anthocyanins followed by the Total flavonoids was calculated as quercetin equivalents (mg QE/100 g DW).
d
same letter are not significantly different at P < 0.05 (Tukey HSD test). Antioxidant activities were calculated as trolox equivalents (mg TE/100 g DW).
270 Z. Cai et al. / Food Chemistry 197 (2016) 266–272

(P < 0.05), whereas time was not significant (P = 0.663). Addition-
ally, the ANOVA results revealed that the extraction parameter
contributions on the response (i.e., anthocyanin yield) were
temperature > power > ethanol concentration. The range analysis
results were consistent with the ANOVA results. The theoretical
maximum anthocyanin yield at 50 °C, 45-min extraction time,
90% ethanol (v/v), and 200 W had the highest Kij values among
all factor levels.
Anthocyanin yield from ASE ranged from 142.46 to
252.34 mg/100 g DW (Table 4). Based on the ANOVA results, the
main effects were temperature and static cycle (P < 0.05), whereas
static time was not significant (P = 0.168). Furthermore, tempera-
ture had a more significant contribution on anthocyanin yield than
static cycle. The theoretical maximum anthocyanin yield at 90 °C,
two static cycles, and 15-min static time had the highest Kij values
among all factor levels.
3.2. Bioactive compound extraction efficiency from PSP
CE, UAE, and ASE were performed under optimized extraction
parameters to assess the extraction efficiency of anthocyanins,
total phenolics, and total flavonoids from PSPs (Table 5). Antho-
cyanin yield from the extraction methods was 217.58–244.07 mg
cyanidin-3-glucoside equivalents (CGE)/100 g DW; an anthocyanin
yield of 0–663 mg/100 g DW (or 0–210 mg/100 g FW) has been
reported in different PSP genotypes using pressurized liquid

Table 6
Qualitative and quantitative analyses of anthocyanins in purple sweet potatoes extracts using different extraction methods.

Peak No. kmax M/Z Anthocyanins Peak area% of anthocyaninsA

CE UAE ASE
2 523 773, 449, 287 Cy-3-soph-5-glc 4.4 ± 0.0a 4.0 ± 0.1a 2.0 ± 0.0b
3 521 787, 463, 301 Pn-3-soph-5-glc 9.0 ± 0.3a 10.1 ± 0.1a 5.9 ± 0.1b
Total amount of nonacyl anthocyanins 13.4 ± 0.3a 14.1 ± 0.2a 7.9 ± 0.1b
4 523 893, 731, 449, 287 Cy-3-p-hydroxybenzoyl soph-5-glc 4.1 ± 0.0a 4.3 ± 0.1a 4.2 ± 0.1a
5 530 907, 745, 463, 301 Pn-3-p-hydroxybenzoyl soph-5-glc 1.0 ± 0.0a 0.8 ± 0.0a 0.5 ± 0.0b
6 523 949, 787, 449, 287 Cy-3-feruloyl soph-5-glc 2.9 ± 0.0a 1.2 ± 0.0b 0.8 ± 0.0c
7 519 935, 773, 449, 287 Cy-3-caffeoyl soph-5-glc 14.0 ± 0.0a 13.4 ± 0.0a 11.2 ± 0.1a
8 520 963, 801, 463, 301 Pn-3-feruloyl soph-5-glc 6.2 ± 0.1a 6.5 ± 0.1a 3.7 ± 0.0b
10 520 949, 787, 463, 301 Pn-3-caffeoyl soph-5-glc 1.7 ± 0.1a 1.9 ± 0.0a 2.4 ± 0.1b
Total amount of monoacyl anthocyanins 29.9 ± 0.2a 28.1 ± 0.1a 22.8 ± 0.3b
9 521 1055, 893, 449, 287 Cy-3-caffeoyl-p-hydroxybenzoyl soph-5-glc 9.7 ± 0.0a 10.6 ± 0.1a 15.8 ± 0.0b
11 523 1111, 949, 449, 287 Cy-3-caffeoyl-feruloyl soph-5-glc 1.3 ± 0.0a 1.4 ± 0.0a 1.4 ± 0.0a
12 521 1069, 907, 463, 301 Pn-3-caffeoyl-p-hydroxybenzoyl soph-5-glc 28.1 ± 0.0a 28.9 ± 0.0a 24.8 ± 0.0a
13 520 1125, 963, 463, 301 Pn-3-caffeoyl-feruloyl soph-5-glc 17.6 ± 0.0a 16.9 ± 0.0a 27.2 ± 0.0b
Total amount of diacyl anthocyanins 56.7 ± 0.0a 57.8 ± 0.1a 86.8 ± 0.0b

Abbreviation: cy, cyanidin; soph, sophoroside; glc, glucoside; pn, peonidin. Values are represented as means ± SD (n = 3). Within each line, means with the same letter are not
significantly different (P 6 0.05).
A
Calculation was based on percentage of peak areas in HPLC chromatogram.
 Z. Cai et al. / Food Chemistry 197 (2016) 266–272
extraction (Truong et al., 2010). In this study, the concentration of
anthocyanins extracted from PSPs was higher than that reported in
previous studies (Montilla, Hillebrand, & Winterhalter, 2011;
Rodriguez-Saona, Millar, & Trumble, 1998; Teow et al., 2007):
32.2–53.1 mg CGE/100 g FW (or 107.32–176.98 mg CGE/100 g
DW; Teow et al., 2007) and 6.5–29.1 mg/100 g FW (Montilla
et al., 2011).
Total phenolic and flavonoid yield was 630.67–955.05 mg gallic
acid equivalents (GAE)/100 g and 27.81–40.54 mg quercetin equiv-
alents (QE)/100 g based on dry mass, respectively (Table 5). In this
study, the concentration of total phenolics was higher than that
obtained by Teow et al. (2007) by the CE method (21.2–42.2 mg
GAE/100 g FW, or 70.66–140.65 mg GAE/100 g DW). Differences
in bioactive components might be attributed to differences in
PSP cultivars or extraction assays.
Interestingly, anthocyanin yield followed the order ASE > UAE >
CE, which was opposite to the total phenolic and flavonoid yield
(CE > UAE > ASE; P < 0.05). A short-time extraction method at high
temperatures (e.g., ASE) might favor the extraction of anthocyanins
which are heat-labile, whereas long-time extraction methods
at relatively low temperatures (e.g., CE and UAE) contribute to
low anthocyanin yields. Other thermo-stable phenolics could be
extracted from PSPs with long extraction durations.
3.3. Antioxidant activity of PSP extracts
Extraction conditions affect antioxidant activities of plant com-
pounds (Michiels, Kevers, Pincemail, Defraigne, & Dommes, 2012).
The ORAC value of CE extracts was 1091.06 mg trolox equivalents
(TE)/100 g DW, similar to that of UAE extracts (1036.94 mg
TE/100 g DW), but higher than that of ASE extracts (766.13 mg
TE/100 g DW; P = 0.01; Table 5). The ORAC values obtained in this
study were higher than those obtained by Teow et al. (2007), who
reported 68.08–83.35 mg TE/100 g FW (or 226.71–277.56 mg
TE/100 g DW) for sweet potato cultivars containing 70% water.
However, the FRAP value of ASE extracts was 1704.76 mg
TE/100 g DW, which was significantly higher than that of CE and
UAE extracts (P < 0.01). PSP anthocyanins had higher FRAP values,
whereas other flavonoids or phenolics had higher ORAC values.
3.4. PSP anthocyanin composition
PSP extracts were subjected to HPLC, UV–visible, and LC–MS. A
total of 13 peaks of PSP anthocyanin-rich extracts had a maximum
absorbance at 519–530 nm. On the other hand, 27 anthocyanin
peaks have been detected in purple-fleshed sweet potatoes grown
in Korea (Lee, Park, Choi, & Jung, 2013). Only 12 anthocyanins
(Table 6) were identified by LC–MS based on fragmentation
patterns of individual peaks and past studies (Kim et al., 2012;
Lee et al., 2013). Peonidin-3-sophoroside-5-glucoside and cyani
din-3-sophoroside-5-glucoside represented 7.9–13.4% of the total
anthocyanins and were identified as the basic structure of other
acylated anthocyanins. Some acylated anthocyanins with one or
two p-hydroxybenzoic, ferulic and/or caffeic acid were monoacyl
(22.8–29.9%) and diacyl anthocyanins (56.7–86.8%), respectively.
The extraction methods affected the peak area percentage of
anthocyanins (Table 6). In the ASE extracts, the fraction of nonacyl,
monoacyl, and diacyl anthocyanins in the 12 anthocyanin peaks
were 7.9%, 22.8%, and 86.8% (peak area percentage), respectively.
ASE extracts contained more diacyl anthocyanins and less nonacyl
and monoacyl anthocyanins than CE or UEA extracts (P < 0.05),
which could be attributed to the thermal degradation effects of
CE and UAE on acyl anthocyanins (Abdel-Aal et al., 2014). ASE
extracts were rich in acylated anthocyanins, which are more stable
than those present in CE and UAE extracts (Giusti & Wrolstad,
2003).
271
4. Conclusions
Extraction methods affect the yield, purity, and composition of
anthocyanins. This study was aimed to compare CE, UAE, and ASE
as to the effectiveness of extraction of anthocyanins from purple
sweet potatoes. While requiring high temp, ASE was a pressured-
liquid and short-time extraction method and thus favored the
extraction of anthocyanins that are heat labile from purple sweet
potatoes. CE and UAE are long-time extraction methods at rela-
tively low temperature and resulted in low anthocyanin yields
with high impurities of other phenolics in this study. As far as
the aspect of anthocyanin composition is concerned, ASE could
extract more diacyl anthocyanins and less nonacyl and monoacyl
anthocyanins than CE or UEA, suggesting anthocyanins in ASE
extracts were more stable than those extracted by CE or UAE. These
methods also affected antioxidant activities of PSPs extracts. When
applied to food matrix as antioxidant agents, PSPs anthocyanin-
rich extracts via the three methods demonstrated different activi-
ties depending on the mechanisms involved. CE and UAE extracts
were more efficient in breaking radical chain activity while ASE
would be more suitable as an antioxidant via the mechanism of
ferric reduction. The anthocyanin extraction method from PSP
should take into consideration the extraction efficiency of bioactive
compounds, antioxidant activities, and types of anthocyanins.
Acknowledgements
This study was founded by National Nature Science Foundation
of China (Grant No. 31371756) and Hainan Provincial department
of Science and Technology (ZDXM2014008).
References
Abdel-Aal, E. S. M., Akhtar, H., Rabalski, I., & Bryan, M. (2014). Accelerated,
microwave-assisted, and conventional solvent extraction methods affect
anthocyanin composition from colored grains. Journal of Food Science, 79(2),
C138–C146.
Carrera, C., Ruiz-Rodriguez, A., Palma, M., & Barroso, C. G. (2012). Ultrasound
assisted extraction of phenolic compounds from grapes. Analytica Chimica Acta,
732, 100–104.
Fan, G. J., Han, Y. B., Gu, Z. X., & Chen, D. M. (2008). Optimizing conditions for
anthocyanins extraction from purple sweet potato using response surface
methodology (RSM). LWT-Food Science and Technology, 41(1), 155–160.
Giusti, M. M., & Jing, P. (2007). Natural pigments of berry fruits: functionality and
application. In Y. Zhao (Ed.), Berry fruit: Value-added products for health
promotion (1st ed., pp. 105–146). Florida: Taylor & Francis.
Giusti, M. M., & Wrolstad, R. E. (2003). Acylated anthocyanins from edible sources
and their applications in food systems. Biochemical Engineering Journal, 14,
217–225.
Giusti, M. M., & Wrolstad, R. E. (2001). Characterization and measurement of
anthocyanins by UV–visible spectroscopy. In R. E. Wrolstad, T. E. Acree, E. A.
Decker, M. H. Penner, D. S. Reid, S. J. Schwartz, C. F. Shoemaker, D. Smith, & P.
Sporns (Eds.), Current protocols in food analytical chemistry (1st ed.,
pp. F1.2.1–F1.2.13). NY: John Wiley & Sons Inc.. Vol. 1.
Hwang, Y. P., Choi, J. H., Yun, H. J., Han, E. H., Kim, H. G., Kim, J. Y., ... Jeong, H. G.
(2011). Anthocyanins from purple sweet potato attenuate
dimethylnitrosamine-induced liver injury in rats by inducing Nrf2-mediated
antioxidant enzymes and reducing COX-2 and iNOS expression. Food and
Chemical Toxicology, 49(1), 93–99.
Jing, P., & Giusti, M. M. (2007). Effects of extraction conditions on improving the
yield and quality of an anthocyanin-rich purple corn (Zea mays L.) color extract.
Journal of Food Science, 72(7), C363–C368.
Jing, P., Ye, T., Shi, H. M., Sheng, Y., Slavin, M., Gao, B. Y., ... Yu, L. L. (2012).
Antioxidant properties and phytochemical composition of China-grown
pomegranate seeds. Food Chemistry, 132(3), 1457–1464.
Jing, P., Zhao, S. J., Ruan, S. Y., Sui, Z. Q., Chen, L. H., Jiang, L. L., & Qian, B. J. (2014).
Quantitative studies on structure-ORAC relationships of anthocyanins from
eggplant and radish using 3D-QSAR. Food Chemistry, 145, 365–371.
Jing, P., Zhao, S. J., Ruan, S. Y., Xie, Z. H., Dong, Y., & Yu, L. L. (2012). Anthocyanin and
glucosinolate occurrences in the roots of Chinese red radish (Raphanus sativus
L.), and their stability to heat and pH. Food Chemistry, 133(4), 1569–1576.
Kim, H. W., Kim, J. B., Cho, S. M., Chung, M. N., Lee, Y. M., Chu, S. M., ... Lee, D. J.
(2012). Anthocyanin changes in the Korean purple-fleshed sweet potato,
Shinzami, as affected by steaming and baking. Food Chemistry, 130(4), 966–972.
Lee, M. J., Park, J. S., Choi, D. S., & Jung, M. Y. (2013). Characterization and
quantitation of anthocyanins in purple-fleshed sweet potatoes cultivated in
272 Z. Cai et al. / Food Chemistry 197 (2016) 266–272
Korea by HPLC-DAD and HPLC-ESI-QTOF-MS/MS. Journal of Agricultural and Food
Chemistry, 61(12), 3148–3158.
Lien, C. Y., Chan, C. F., Lai, Y. C., Huang, C. L., & Liao, W. C. (2012). Ultrasound-assisted
anthocyanin extraction of purple sweet potato variety TNG73, Ipomoea batatas,
L.. Separation Science and Technology, 47(8), 1241–1247.
Mazza, G., & Miniati, E. (1993). Anthocyanins in fruits, vegetables and grains London.
CRC Press.
Meda, A., Lamien, C. E., Romito, M., Millogo, J., & Nacoulma, O. G. (2005).
Determination of the total phenolic, flavonoid and proline contents in Burkina
Fasan honey, as well as their radical scavenging activity. Food Chemistry, 91(3),
571–577.
Michiels, J. A., Kevers, C., Pincemail, J., Defraigne, J. O., & Dommes, J. (2012).
Extraction conditions can greatly influence antioxidant capacity assays in plant
food matrices. Food Chemistry, 130(4), 986–993.
Montilla, E. C., Hillebrand, S., & Winterhalter, P. (2011). Anthocyanins in purple
sweet potato (Ipomoea batatas L.) varieties. In J. A. T. D. Silva (Ed.), Fruit vegetable
and cereal science and biotechnology (pp. 19–24). Global Science Books.
Rao, R. S., Kumar, C. G., Prakasham, R. S., & Hobbs, P. J. (2008). The Taguchi
methodology as a statistical tool for biotechnological applications: A critical
appraisal. Biotechnology Journal, 3, 510–523.
Rastogi, N. K. (2011). Opportunities and challenges in application of ultrasound in
food processing. Critical Reviews in Food Science and Nutrition, 51(8), 705–722.
Rodriguez-Saona, C., Millar, J. G., & Trumble, J. T. (1998). Isolation, identification,
and biological activity of isopersin, a new compound from avocado idioblast oil
cells. Journal of Natural Products, 61(9), 1168–1170.
Rodriguez-Solana, R., Salgado, J. M., Dominguez, J. M., & Cortes-Dieguez, S. (2015).
Comparison of soxhlet, accelerated solvent and supercritical fluid extraction
techniques for volatile (GC–MS and GC/FID) and phenolic compounds (HPLC-
ESI/MS/MS) from lamiaceae species. Phytochemical Analysis, 26(1), 61–71.
Santos, D. T., Veggi, P. C., & Meireles, M. A. A. (2010). Extraction of antioxidant
compounds from Jabuticaba (Myrciaria cauliflora) skins: Yield, composition and
economical evaluation. Journal of Food Engineering, 101(1), 23–31.
Teow, C. C., Truong, V. D., McFeeters, R. F., Thompson, R. L., Pecota, K. V., & Yencho, G.
C. (2007). Antioxidant activities, phenolic and beta-carotene contents of
sweet potato genotypes with varying flesh colours. Food Chemistry, 103(3),
829–838.
Truong, V. D., Deighton, N., Thompson, R. T., McFeeters, R. F., Dean, L. O., Pecota, K.
V., & Yencho, G. C. (2010). Characterization of anthocyanins and anthocyanidins
in purple-fleshed sweet potatoes by HPLC-DAD/ESI-MS/MS. Journal of
Agricultural and Food Chemistry, 58(1), 404–410.
Truong, V. D., Hu, Z., Thompson, R. L., Yencho, G. C., & Pecota, K. V. (2012).
Pressurized liquid extraction and quantification of anthocyanins in
purple-fleshed sweet potato genotypes. Journal of Food Composition and
Analysis, 26(1–2), 96–103.
Waterhouse, A. L. (2001). Determination of total phenolics. In R. E. Wrolstad, T. E.
Acree, H. An, E. A. Decker, M. H. Penner, D. S. Reid, S. J. Schwartz, C. F.
Shoemaker, & P. Sporns (Eds.), Current protocols in food analytical chemistry (1st
ed., pp. I1.1.1–11.11.12). NY: John Wiley & Sons Inc..
Wrolstad, R. E., & Culver, C. A. (2012). Alternatives to those artificial FD&C food
colorants. Annual Review of Food Science and Technology, 3(1), 59–77.
Zhang, Z. F., Fan, S. H., Zheng, Y. L., Lu, J., Wu, D. M., Shan, Q., & Hu, B. (2009). Purple
sweet potato color attenuates oxidative stress and inflammatory response
induced by D-galactose in mouse liver. Food and Chemical Toxicology, 47(2),
496–501.