Accepted Manuscript

Title: Extraction and purification of anthocyanins from

purple-fleshed potato

Author: Jari Heinonen Hengameh Farahmandazad Anssi

Vuorinen Heikki Kallio Baoru Yang Tuomo Sainio

PII: S0960-3085(16)30036-0
DOI: http://dx.doi.org/doi:10.1016/j.fbp.2016.05.004
Reference: FBP 716

To appear in: Food and Bioproducts Processing

Received date: 23-12-2015

Revised date: 2-5-2016
Accepted date: 9-5-2016

Please cite this article as: Heinonen, J., Farahmandazad, H., Vuorinen, A.,
Kallio, H., Yang, B., Sainio, T.,Extraction and purification of anthocyanins
from purple-fleshed potato, Food and Bioproducts Processing (2016),
http://dx.doi.org/10.1016/j.fbp.2016.05.004

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
HIGHLIGHTS

- A process for recovering anthocyanins from purple-fleshed potato is presented.

- Anthocyanins are extracted from potato tubers using aqueous ethanol.

t
- Purification of the anthocyanins is done efficiently using XAD-7HP as adsorbent.

ip
- Adsorption removes practically all impurities from the extracts.
- The process can be easily scaled up and thus used on a larger scale.

cr
us
an
M
ed
pt
ce
Ac

Page 1 of 37
 t
ip
cr
us
an
Extraction and purification of anthocyanins from purple-fleshed potato

Jari Heinonen*,1, Hengameh Farahmandazad1, Anssi Vuorinen2,

M
Heikki Kallio2, Baoru Yang2, Tuomo Sainio1

1
School of Engineering Science
Lappeenranta University of Technology
ed

Skinnarilankatu 34, FIN-53850 Lappeenranta, Finland

2
Food Chemistry and Food Development
pt

Department of Biochemistry
University of Turku
Vatselankatu 2, FIN-20014 Turku, Finland
ce
Ac

* Corresponding author. Tel: +358-40-127 2920, e-mail: jari.heinonen@lut.fi

ABSTRACT
2

Page 2 of 37
Extraction and purification of anthocyanins from purple-fleshed potatoes was investigated. The
effects of pretreatment method, time, and ethanol concentration on the extraction were studied.
Boiling was the most efficient pretreatment method for the extraction. Extraction time and
ethanol concentration were found to have only minor effect on the extraction yield. Six hour
extraction time and 20 vol.% ethanol concentration were found to be sufficient. The resulting
anthocyanin-rich extract contained glucose, proteins, amino acids, and unidentified compounds as
impurities. Removal of these by adsorption was studied with six commercially available

t
adsorbents, of which Amberlite XAD-7HP was the most efficient. Anthocyanin purification with

ip
XAD-7HP was studied in a column. The effects of flow rate on adsorption, and ethanol and
acetic acid concentrations on desorption were investigated. In loading, a highly dispersed

cr
anthocyanins breakthrough curve was obtained. This most likely resulted from agglomeration of
anthocyanins. Desorption of anthocyanins could be done efficiently with 75 vol.% acidic
(7 vol.% acetic acid) aqueous ethanol. All glucose and unknown impurities, together with 87 %

us
of proteins and amino acids, could be separated from the anthocyanins. Process scale-up was
demonstrated by processing 19 kg of purple-fleshed potatoes in a single batch.

an
M
ed
pt

Keywords: purple-fleshed potato, anthocyanins, extraction, adsorption, purification

ce
Ac

Page 3 of 37
NOTATION
A700 absorbance of anthocyanins at the UV wavelength of 700 nm, -
Amax absorbance of anthocyanins at the UV wavelength giving the maximum absorbance, -

t
Canthocyanin anthocyanin concentration, g/L

ip
hbed length of the adsorbent bed, cm
L path length of the quartz cuvette in UV detector, cm
M molecular weight, g/mol

cr
madsorbent mass of the adsorbent, g
n sample size, -
p pressure, atm

us
SEx standard error of mean, mg/g or mg/L
T temperature, °C
V volume, L
Vdesorbent
Vextract
xi
volume of the desorbent solution, L an
volume of the extract in contact with the adsorbent, L
value of sample i, mg/g or mg/L
x sample mean value, mg/g or mg/L
M
Greek letters
δ dilution factor
ed

ε molar extinction coefficient, L/(mol cm)

Subscripts and superscripts

des desorption step
pt

ads adsorption step

eq in equilibrium
0 initial
ce

Acronyms
BV bed volume
Ac

Page 4 of 37
1 INTRODUCTION

Polyphenols are secondary plant metabolites which have been shown to have positive effects on
human health. Plants contain many types of polyphenols such as anthocyanins which are natural

t
pigments seen as red, blue, and purple colors in flowers, fruits, and vegetables [1]. Anthocyanins

ip
are antioxidants that scavenge free radicals and reduce oxidative stress by donating protons to
highly reactive radicals [2,3]. They also possess antimicrobial, antiviral, and anti-inflammatory

cr
properties [4,5]. Long-term dietary intake of anthocyanins may reduce the risk of cancer,
cardiovascular diseases, virus infections, Alzheimer’s disease and diabetes [6]. In addition, their

us
use has been shown to improve night vision [7]. They are also used as natural pigments by the
food and beverage industries, as well as in cosmetics and paint sector [6,8].
an
Anthocyanins are present in many plants and can be obtained from them by means of extraction.
Different extraction methods have been investigated, including conventional solid–liquid
M
extraction (e.g., [5,9]), supercritical fluid extraction (e.g., [6]), and ultrasound assisted extraction
(e.g., [6,10]). Out of these, the solid–liquid extraction with an aqueous organic solvent has so far
been the most popular method. Burgos et al. [5] used 60 vol.% and 80 vol.% aqueous methanol to
ed

extract anthocyanins from raw and boiled potato (Solanum tuberosum L.) tubers. 60 vol.%
methanol was found to give a higher yield for raw potatoes, while 80 vol.% was most efficient
pt

with the boiled potatoes [5]. Bridgers et al. [9] reported that aqueous acidic methanol gave a
higher anthocyanin yield from purple fleshed sweet potatoes than aqueous acidic ethanol. In
ce

extraction of anthocyanins from black rice, methanol (70 vol.%) yielded a higher anthocyanin
concentration than ethanol (70 vol.%), acetone (70 vol.%), or water [11]. In another study, from
the various extraction media tested, Chandrasekhar et al. [12] obtained the highest anthocyanin
Ac

yield from red cabbage using acidic 50 vol.% aqueous ethanol. Patil et al. [13] tested the effect of
ethanol concentration on the extraction yield, and reported that the highest yield was obtained
with aqueous acidic 50 vol.% ethanol. As seen from above, both methanol and ethanol have been
successfully used in the extraction. However, ethanol is preferred due to its low toxicity.

In most studies dealing with the extraction of anthocyanins, addition of a small amount of acid to
the extraction medium has been reported to increase anthocyanin yield. This is caused by
stabilization of anthocyanins at low pH. However, too high acid concentration leads to hydrolysis
5

Page 5 of 37
of acylated anthocyanins; consequently, a decrease in pH to approximately three by acid addition
provides favorable conditions for extraction as, at this pH, anthocyanins are in a stable flavylium
ion form [14]. Different acids such as hydrochloric, malic, tartaric, ascorbic, citric, and acetic
acid have been used to adjust pH [4,9-12,15-18]. Of these, acetic acid is the most suitable when
anthocyanins are to be used in the food and beverage industries. Truong et al. [16] have shown

t
ip
that 5-10 vol.% acetic acid in the extraction medium is the optimal concentration for anthocyanin
extraction.

cr
In addition to anthocyanins, the extraction process releases other compounds from plants,

us
including free sugars, sugar alcohols, organic acids, amino acids, and proteins [12,19]. Some of
these, such as free sugars, cause degradation of anthocyanins during storage or hinder the
downstream processing [12,20]. Thus, separation of anthocyanins from these compounds is
required.
an
Purification of anthocyanins can be done by adsorption which is an efficient and simple method.
M
Adsorptive purification of anthocyanin-rich extracts has been carried out quite extensively.
Chandrasekhar et al. [12] studied the use of silica gel, Amberlite IRC 80, Amberlite IR 120,
ed

DOWEX 50WX8, Amberlite XAD-4, and Amberlite XAD-7HP for the purification of
anthocyanins from red cabbage extract. Amberlite XAD-7HP was found to be the most efficient.
Desorption of adsorbed anthocyanins was conducted successfully with acidic aqueous ethanol of
pt

a concentration higher than 60 vol.% [12]. Use of the same adsorbents was investigated by
Jampani et al. [19] for the purification of anthocyanins from jamun (Syzygium cumini) extract.
ce

Again, XAD-7HP was found to be the most efficient, and desorption of anthocyanins could be
carried out using acidic aqueous ethanol (≥ 40 vol.%) [12]. Investigation of the recovery of
Ac

anthocyanins from black rice by Kang et al. [11] showed that out of XAD-4, XAD-7HP, HP20,
HP2MG, SP70, and SP207 adsorbents, the highest adsorption capacity was achieved with XAD-
7HP. Efficient elution of the bound anthocyanins was carried out with acidic aqueous 60 vol.%
ethanol [11]. XAD-7HP has also been successfully applied in the purification of anthocyanins
from red cabbage juice by Coutinho et al. [21]. Wang et al. [22] investigated the purification of
anthocyanins from black bean canning water using five macroporous polystyrene based resins.
Highest purification efficiency was obtained with the resin having the largest surface area,
Sepabeads Sp700 [22]. Liu et al. [23] studied the recovery of anthocyanins from purple-fleshed
6

Page 6 of 37
potatoes. Purification of anthocyanins was carried out by membrane filtration and adsorption with
eleven adsorbents. Amberlite XAD-1600 was found to be the most efficient [23].

Although extraction and adsorptive purification of anthocyanins have been studied extensively,
there is still a paucity of information related to the extraction and purification of anthocyanins

t
from purple-fleshed potatoes. To the best of our knowledge, the only reports available

ip
[5,10,18,23] do not cover aspects such as extraction conditions with aqueous ethanol, desorption

cr
of anthocyanins, the behavior of the impurities during the purification, column operations, or
process scale-up. Anthocyanins in purple-fleshed potatoes are highly valuable as they are mostly

us
acylated glycosides of petunidin and peonidin; for example, petunidin-coumaroylrutinoside-
glucoside and peonidin-p-coumaroylrutinoside-glucoside [24,25]. Acylated anthocyanins are
more stable than nonacylated anthocyanins. Potatoes are also a cheap and readily available raw
an
material. Accordingly, the aim of this work was to present an efficient method for obtaining
anthocyanins from the Finnish purple-fleshed potato landrace cultivar Synkeä Sakari. The
M
extraction and adsorptive purification of these anthocyanins were investigated. The main
emphasis was on the adsorptive purification as the extraction of anthocyanins from various raw
materials has been investigated extensively [5, 9,11-18]. Thus, only a limited study of extraction
ed

was conducted here to find extraction conditions in which sufficient anthocyanin yield could be
obtained. In the adsorptive purification, six commercially available adsorbents were compared by
pt

batch equilibrium experiments. The adsorption was also studied by column experiments with the
most efficient adsorbent. In addition, scale-up of the extraction and purification process was
ce

studied by processing 19 kg of purple-fleshed potatoes in a single batch.

Ac

2 EXPERIMENTS

2.1 Materials

Anthocyanins were extracted and purified from tubers of a Finnish cultivar of purple-fleshed
potato, Synkeä Sakari. The potatoes were cultivated in Kokemäki, Finland; from May 2014 to
September 2014 (duration: 132 days). The harvested potatoes were stored at a temperature of

Page 7 of 37
4 °C. Two batches of potatoes were used: first batch of completely purple-fleshed potatoes, and
second batch of partially pigmented potatoes.

Analysis grade ethanol (99.9 % VWR International), acetic acid (100 % Merck KGaA), sodium
chloride (99.7 % VWR International), potassium chloride (99.5 % Riedel-de Haën), and sodium

t
acetate (99 % Riedel-de Haën) were used in the experiments. Purified water (purification using

ip
CENTRA-R 60/120, ELGA/VEOLIA water) was used in all experiments.

cr
2.2 Pretreatment of potato tubers

us
Four different pretreatment methods were tested: freezing of raw tubers, freeze-drying of raw
tubers, boiling of raw tubers and subsequent freezing, and boiling and drying of raw tubers with
subsequent freezing. Each experiment was repeated three times to obtain reliable results with
an
standard errors to show the experimental inaccuracy. The standard error of mean SEx was
calculated with
M
n

 x  x
2
i
1
SE x  i 1
, (1)
ed

n n 1

where n is the sample size, xi is the sample value, and x is the sample mean value.
pt

Tubers were cut into 0.5 cm thick pieces before freezing or freeze-drying. For freezing, tuber
ce

pieces were stored in a freezer at a temperature of ˗20 °C. After 24 hours, the pieces were ground.

For freeze-drying, tuber pieces were placed in a freezer (˗20 °C) for three days. Afterwards, the
Ac

frozen pieces were put into a freeze-dryer (T = ˗76 °C, p = 1 bar, Christ Alpha 2-4 LSC Freeze
Dryer) for four days. The freeze-dried samples were ground, and stored in a freezer at ˗20 °C.

Boiled tubers were obtained by boiling whole tubers in water for 25 minutes and then smashed.
Part of the smashed tubers were stored in a freezer (˗20 °C), and part were baked in an oven at
110 °C for one hour. The baked tubers were stored in a freezer (˗20 °C).

Page 8 of 37
2.3 Extraction of anthocyanins

Anthocyanins were extracted with conventional extraction method. The aim was to use them in
clinical trials; accordingly, the extraction was carried out using aqueous ethanol. The effect of
ethanol concentration on the extraction was studied with 20, 40, and 80 vol.% aqueous ethanol

t
solutions. Each experiment was repeated three times to obtain reliable results with standard errors

ip
(calculated with Eq. (1)) to show the experimental inaccuracy.

cr
In order to prevent the degradation of nonacylated anthocyanins and to keep the extracted
anthocyanins in a stable form (as flavylium ions), pH of the extraction medium was kept mildly

us
acidic using acetic acid (7 vol. %).

The extraction was studied with a batch reactor at room temperature. First, the pretreated potatoes
an
(appr. 10-100 g depending on the experiment) were added to the extraction medium with a solid–
liquid ratio of 1:5 (wt.:wt.) and mixed for 15 minutes. Afterwards, the samples were kept in dark
M
up to 24 hours. The resulting extracts were filtered under a vacuum to remove coarse particles.
Centrifuge (4 000 rpm, 20 minutes; Heraeus Megafuge 1.0) was used to remove the fine solids.
The solid-free extracts were stored in a freezer (˗20 °C).
ed

2.4 Ethanol removal and concentration of the extract

pt

Two methods were used in this work to remove ethanol and to concentrate the anthocyanin
extracts: evaporation under vacuum and evaporation under atmospheric pressure. In evaporation
ce

under vacuum, a rotary evaporator (Heidolph Hei-VAP Advantage Rotary Evaporator) was used
at a temperature of 35 °C with 60 mbar pressure. Evaporation under atmospheric pressure was
Ac

done at a temperature of 60 °C. The concentrated extracts were stored in a freezer (˗20 °C).

2.5 Purification of anthocyanins by adsorption

2.5.1 Screening of suitable adsorbents

Six commercially available adsorbents (Table 1) were tested for the purification of anthocyanins
by batch experiments. Each experiment was repeated three times to obtain reliable results with
standard errors (calculated with Eq. (1)) to show the experimental inaccuracy.

Page 9 of 37
<<Table 1 around here>>

Prior to use, the polymeric adsorbents were washed with 5 BV (bed volumes) of 20 vol.% ethanol
and 5 BV of purified water to remove preservatives. The cation exchange resins, Amberlyst 15
and CS16GC, were converted to Na+ form with 1 mol/L NaCl. Purified water was used to remove

t
excess NaCl. Excess water in each adsorbent was removed by centrifugation (2 600 rpm, 15 min;

ip
Heraeus Megafuge 1.0).

cr
Batch equilibrium experiments were carried out at room temperature to compare the adsorbents.
4 g of each adsorbent was mixed with 16 g of ethanol-free anthocyanin extract. The batches were

us
equilibrated for 15 hours at room temperature. The resulting liquid phase was separated from the
solids by filtration under vacuum, and the anthocyanin concentration in the liquid phase was
determined. an
The adsorbed amount of anthocyanins qanthocyanin was calculated using
M
qanthocyanin 
C
ads
anthocyanin,0  Canthocyanin,eq
ads
Vextract , (2)
madsorbent
ed

ads
where Canthocyanin,0 ads
is the initial anthocyanin concentration in the extract, Canthocyanin,eq is the anthocyanin
pt

concentration in the extract in equilibrium, Vextract is the volume of the extract in contact with the
adsorbent, and madsorbent is the mass of the adsorbent.
ce

Desorption of adsorbed anthocyanins was studied with XAD-4, XAD-7HP, XAD-16N, and MN-
200. Adsorbents loaded with anthocyanins were washed with 3 BV of purified water to remove
Ac

excess anthocyanins. Excess water was removed by centrifugation. Each loaded adsorbent was
contacted with 14.6 g of 50 vol. % ethanol for one hour at room temperature. The resulting liquid
phase was analyzed for anthocyanins. The desorption ratio was calculated with

des
Canthocyanin,eqVdesorbent
Desorption ratio   100% , (3)
C ads
anthocyanin,0  Canthocyanin,eq
ads
Vextract

10

Page 10 of 37
where Ceqdes is the anthocyanin concentration in desorption step in equilibrium and Vdesorbent is the

volume of the desorbent solution.

2.5.2 Dynamic purification experiments

t
Purification of anthocyanins extracted from purple-fleshed potatoes was investigated with XAD-

ip
7HP in a column at room temperature. The experimental setup consisted of a peristaltic pump
(Ismatec IP-8, IDEX Health & Science) for liquid feeding, a chromatographic column (ECO SR

cr
15/200, YMC Europe) with 15 mm inner diameter, a UV-Vis spectrophotometer (8453 UV-Vis
spectroscopy system, Agilent) with a flow-through cell (1 cm flow path length), and a fraction

us
collector (Frac-920, GE Healthcare Bio-Sciences). The volume of the adsorbent bed was 17.7 mL
(hbed = 10 cm). Anthocyanin concentration in the feed was 224.5 mg/L.
an
The effect of flow rate (1.7–10.2 BV/h; BV = bed volume) on the adsorption of anthocyanins was
investigated. Washing of the loaded resin was done with purified water (3.4 BV/h) to remove the
M
non-adsorbed solutes from the adsorbent bed.

Desorption of anthocyanins from the loaded adsorbent was performed using aqueous ethanol
ed

solutions. The effect of ethanol concentration (25–75 vol.%) on desorption was investigated. In
addition, the effect of acetic acid addition (7 vol.%) to the eluent was studied. The adsorbent used
pt

in desorption experiments was loaded in a batch reactor in order to have the same amount of
anthocyanins in the adsorbent bed. Washing of the loaded resin with purified water was done
ce

prior to desorption.

2.6 Scale up of the extraction and purification process

Ac

The scalability of the extraction and purification process was demonstrated by producing purified
anthocyanins from 19 kg of completely pigmented potatoes in a single batch. Freeze-drying was
used as the pretreatment method. The extraction of anthocyanins was carried out using aqueous
20 vol.% acidic (7 vol.% acetic acid) ethanol solution with solid:liquid ratio of 1:5. Coarse solids
were separated from the resulting extract by filtration under vacuum, and the fine solids by
centrifugation (2 600 rpm, 15 min, ALC 4237, ALC International). The coarse particles from the
first extraction were contacted again with the extraction medium with solid:liquid ratio of 1:3 in

11

Page 11 of 37
order to maximize the anthocyanin yield. Removal of ethanol from the solid-free extract was
done by evaporation (p = 1 atm, T = 60 °C).

Purification of anthocyanins was carried out in a 1.83 L glass column (hbed = 93 cm; ECO SR
50/999, YMC Europe) at room temperature with XAD-16 as an adsorbent. This adsorbent was

t
used instead of XAD-7HP due to better availability. The flow rate was 0.75 BV/h (pump: Series

ip
III, Labhut). In the loading step, 11.5 L of raw extract was fed to the column. After the loading

cr
step, the adsorbent bed was washed with 3 BV of purified water. Desorption of anthocyanins was
done with 75 vol.% acidic (7 vol.% acetic acid) ethanol solution. Removal of ethanol from the

us
purified anthocyanin-rich solution was done by evaporation (p = 1 atm, T = 60 °C).

2.7 Chemical analyses

2.7.1 Quantification of anthocyanins

an
The total anthocyanin concentration was determined using the pH-differential method [19,26,27].
M
The anthocyanin-rich samples were diluted with 0.025 mol/L potassium chloride buffer (pH 1.0)
and 0.4 mol/L sodium acetate buffer (pH 4.5), and equilibrated for 20 minutes. To determine the
ed

appropriate dilution factor, the samples were first diluted with pH 1.0 buffer until the absorbance
at wavelengths between 520 and 530 nm was within the linear range of the spectrophotometer
(Agilent 8453 UV-Vis Spectrophotometer).
pt

The UV wavelength giving the maximum absorbance with anthocyanins is in the range of 520–
ce

530 nm depending on the type and source of anthocyanins [19,26,27]. The maximum with
anthocyanins from purple-fleshed potatoes was 524 nm.
Ac

The UV absorbances of the anthocyanin samples diluted with pH 1.0 and 4.5 buffer solutions
were then measured with wavelength 524 nm, and also with 700 nm to correct the background
noise. The total anthocyanin concentration Canthocyanin (as cyanidin-3-glucoside equivalents) [g/L]
was calculated from [19,26,27]

A M 
Canthocyanin  (4A)
 L

12

Page 12 of 37
 A   Amax  A700  pH 1 .0   Amax  A700  pH  4 .5 (4B)

where Amax is the UV absorbance at the wavelength with the maximum UV absorbance, A700 is
the UV absorbance at a wavelength of 700 nm, M is the molecular weight of cyanidin-3-
glucoside (449.2 g/mol), δ is the dilution factor, ε is the molar extinction coefficient for cyanidin-

t
ip
3-glucoside (26 900 L/(mol cm)), and L is the path length of the quartz cuvette (1 cm).

cr
2.7.2 Quantification of sugars, ethanol, and anthocyanins with HPLC and HPLC-MS

An HPLC (HP/Agilent 1100, Agilent) with a refractive index detector was used to determine

us
glucose and ethanol concentrations in the anthocyanin extracts. A sugar analysis column Shodex
SP-0810 (Waters) was used with a Shodex SP-G (Waters) guard column. Eluent was purified
an
water, flow rate was 0.5 mL/min, temperature was 80 °C, and injection volume was 10 μL.

The type of the extracted anthocyanins was determined with a UPLC (Acquity Ultra Performance
M
LC, Waters) linked with a quadruple mass spectrometer (Quattro Premier, Waters). Separation of
anthocyanins was carried out on a Kinetex C18 column (Phenomenex) with an AJO-8946 guard
column (Phenomenex). Two LC-MS grade mobile phases, 5 % formic acid in water (A) and
ed

acetonitrile (B), were used. The solvent gradient during the analyses was: 0 minutes 5 % B, 4-44
minutes 11 % B, 44.01 minutes 12 % B, 50 minutes 14 % B, 65 minutes 17 % B, 69-72 minutes
pt

80 % B, 76-80 minutes 5 % B. The flow rate was 1.0 mL/min, of which 0.33 mL/min was
diverted to the MS. The injection volume was 10 µL.
ce

ESI-MS analysis for anthocyanins was performed in positive ion mode. The mass spectral
conditions were: capillary voltage 3.25 kV, cone voltage 45 V, extractor voltage 2.50 V, source
Ac

temperature 150 °C, desolvation temperature 400 °C, cone gas flow 47 L/h, desolvation gas flow
799 L/h. The MS analysis (full scan) was acquired at a range of m/z 250-1000.

2.7.3 Molar masses

The average molar masses of the compounds in the extracts were analyzed with an HPLC
(Agilent Infinity 1260, Agilent) combined with a multi-angle laser light scattering (MALLS)
detector (miniDAWN TREOS, Wyatt Technology). A three column system (Acquity APC AQ
45, 125 and 200, Waters) in series was used with 50 mmol/L NaNO3 solution as eluent. Flow rate
13

Page 13 of 37
was 0.1 mL/min and injection volume was 100 µL. A dn/dc value of 0.18 mL/g was used in
calculations. Astra software was used to analyze data.

2.7.4 Total nitrogen

t
The concentrations of proteins and amino acids, as the only nitrogen-containing compounds in

ip
significant quantities [28,29] in the extracts, were determined indirectly as the amount of total
nitrogen using a Shimadzu TOC-L (Shimadzu) total organic carbon analyzer equipped with a

cr
nitrogen analysis unit.

us
3 Results and discussion

3.1 Extraction of anthocyanins

an
Boiling was found to be the most efficient pretreatment method for the extraction (Fig. 1). It
results in gelatinization of starch, which is thought to soften the potatoes. During boiling, the
M
starch granules also swell due to water binding, and the cell walls soften. This results in better
diffusion of the extraction medium into the potato increasing the extraction efficiency.
Conversely, if boiled tubers are dried, water is drawn out and, as a result, their structure changes,
ed

making the extraction less efficient (Fig. 1).

pt

In freeze-drying, water is evaporated from the tubers in vacuum conditions and at low
temperature. The structure of the tubers remains intact, but a porous structure is formed [30].
ce

When freeze-dried tubers are mixed with the extraction medium, the pores are filled with the
solvents. This increases the extraction efficiency. However, freeze-drying was not found to be as
efficient a pretreatment method as boiling (Fig. 1).
Ac

Extraction of anthocyanins from raw tubers was found to be the least efficient pretreatment
method (Fig. 1). In raw tubers, the structure is intact and no porous structure, similar to that seen
in freeze-drying, is formed. Therefore, the diffusion of solvents is slower and the extraction yield
obtained in a reasonable time is low.

<<Figure 1 around here>>

14

Page 14 of 37
Approximately 80 % of anthocyanins that could be extracted from the purple-fleshed potato
tubers over 24 hours were extracted during the first two hours (Fig. 2). However, there is still a
notable increase (larger than the error) in the concentration of anthocyanins between four and six
hours (Fig. 2). Therefore, six hours was deemed a sufficient time to extract anthocyanins from
boiled purple-fleshed potatoes. However, while preparing anthocyanin extracts for the

t
ip
purification experiments, a 24-hour extraction time was used to maximize the anthocyanin yield.

cr
<<Figure 2 around here>>

Increase in ethanol concentration (20–80 vol. %) in the extraction medium enhanced the amount

us
of anthocyanins that could be extracted from boiled purple-fleshed potatoes only slightly (Fig. 3).
However, as concentration of the anthocyanin extract should be carried out after the extraction,
an
higher ethanol concentration in the extract is favorable in this respect as the evaporation of
ethanol is easier than that of water due to the higher vapor pressure of ethanol
(pethanol(20 °C) = 5.8 kPa [37]; pwater(20 °C) = 2.4 kPa [37]).
M
Concentration of the extract and removal of ethanol can be carried out using evaporation (under
vacuum) or, for example, by nanofiltration (e.g., Ref [31]). When considering an industrial scale
ed

process, both of these concentration methods result in relatively low chemical consumption as
ethanol removed from the extracts can be recycled. However, high inflammability of ethanol
pt

would increase the evaporation costs considerably and thus, for example, nanofiltration would be
more suitable for ethanol removal.
ce

In this work, the extracts were concentrated (removal of ethanol) using evaporation both under
vacuum (T = 35 °C) and under atmospheric pressure (T = 60 °C). As relatively high temperature
Ac

was used in the evaporation under atmospheric pressure, thermal degradation of anthocyanins
was expected [38,39,40]. In fact, approximately 11 % decrease in the anthocyanin concentration
was observed. In evaporation under vacuum notable degradation was not observed due to milder
temperature. The extracts used in the adsorption experiments were concentrated by evaporation
under vacuum.

<<Figure 3 around here>>

15

Page 15 of 37
Detailed analysis of the unpurified concentrated anthocyanin extract that was used in the
purification experiments was carried out. The anthocyanin concentration in the extract was
224.5 mg/L. The type of anthocyanins was determined using HPLC-MS analysis (Fig. 4A). The
main species in the extract were identified as acylated anthocyanins petunidin-
coumaroylrutinoside-glucoside and peonidin-p-coumaroylrutinoside-glucoside. No

t
ip
polymerization of anthocyanins [32-35] was detected in molar mass analyses with HPLC-
MALLS. Over 99 % of the mass fraction in the raw extract was caused by molecules with a

cr
molar mass in range of 400-600 g/mol with polydispersity of approximately 1.4 (polydispersity
of one equals molecules with all having the same molar mass). The rest of the mass fraction (less

us
than 1 %) was due to molecules having molar mass above 650 000 g/mol. These were most
probably proteins.
an
HPLC analyses of sugars showed two unknown impurities in the extract in addition to glucose
and ethanol peaks (Fig. 4B). The concentration of glucose in the concentrated raw extract was 3.0
M
g/L.

<<Figure 4 around here>>

ed

The total nitrogen content of the concentrated extract was also measured. The nitrogen content is
related to the protein and amino acid content as these are the only nitrogen containing species in
pt

potato in significant quantities [28,29]. These are also impurities in the anthocyanin extracts and
should be removed. The total nitrogen content of the raw extract was 459 mg/L.
ce

3.2 Purification of anthocyanins extracted from purple-fleshed potato

Ac

3.2.1 Screening of adsorbents

The strongest adsorption of anthocyanins was obtained with the polymeric adsorbents, and the
weakest with the strong cation exchange resins in Na+ form (Fig. 5). Of the polymeric adsorbents,
anthocyanins were most strongly adsorbed to those with PS–DVB matrices. This was due to the
strong affinity between the aromatic anthocyanins and the aromatic resin matrices. The cation
exchange resins also had PS–DVB matrices, but the sulfonic acid functional groups reduced the
affinity considerably. The adsorbed amounts of anthocyanins with the three hydrophobic

16

Page 16 of 37
adsorbents with PS–DVB matrices were almost equal (Fig. 5) although the PS–DVB adsorbents
have different surface areas (see Table 1). This was not due to total adsorption of anthocyanins in
the solution as, at equilibrium, the anthocyanin concentration in the solution was still
approximately 8 % of initial concentrations with each of these adsorbents.

t
XAD-7HP with an acrylic matrix also showed a good adsorption capacity for anthocyanins

ip
(Fig. 5). However, it was slightly lower than with the PS–DVB adsorbents. This was not

cr
expected on the basis of earlier results that [12,19] showed that the more polar matrix of XAD-
7HP should have resulted in stronger adsorption of anthocyanins on this adsorbent than on the

us
PS–DVB adsorbents. However, the surface area of XAD-7HP is smaller than that of the PS–
DVB adsorbents (see Table 1) being a probable reason for the slightly weaker adsorption.

an
Desorption of adsorbed anthocyanins from the polymeric adsorbents was conducted with
50 vol.% aqueous ethanol (Fig. 5). In this case, XAD-7HP performed considerably better than the
others, and 71 % of anthocyanins in the adsorbent could be recovered in the experimental
M
conditions in question. Less than 50 % of anthocyanins could be recovered from the other
adsorbents. The large desorption ratio with XAD-7HP was a result of the hydrophilic acrylic
ed

matrix. Ethanol is more strongly adsorbed to an adsorbent with an acrylic matrix than to an
adsorbent with a PS–DVB matrix. Thus, with XAD-7HP, ethanol displaces the anthocyanins
from the resin more efficiently than from the other adsorbents. On the basis of the high
pt

desorption ratio, XAD-7HP was chosen for further studies.

ce

<<Figure 5 around here>>

3.2.3 Purification of anthocyanins with XAD-7HP in a column

Ac

Breakthrough curves of anthocyanins and glucose in a column packed with XAD-7HP adsorbent
obtained with different flow rates are shown in Fig. 6. The anthocyanin front quickly emerged
from the column after the void volume of the adsorbent bed regardless of flow rate. The
breakthrough curves of anthocyanins were very wide in each case. Possible causes for this are
slow mass transfer and polymerization or agglomeration of anthocyanins. Dispersion caused by
slow mass transfer can be decreased by decreasing flow rate (Fig. 6). However, the breakthrough
curve of anthocyanins was still very wide even with the lowest flow rate (1 BV/h). Therefore, the

17

Page 17 of 37
shape most probably resulted from agglomeration or polymerization [32-35] of anthocyanins in
the extract. These phenomena would prevent the adsorption of anthocyanins on XAD-7HP due to
size exclusion. As a result, these molecules would rapidly propagate through the column and a
wide breakthrough curve, as seen here, would be obtained.

t
Molar mass analyses showed no polymerization of anthocyanins in the samples taken from the

ip
column outlet during the loading. Over 99 % of the mass fraction of this solution was due to

cr
molecules with molar masses in range of 400-600 g/mol; the rest having molar mass larger than
650 000 g/mol. The polydispersity of the small molecules was approximately 1.4 which was the

us
same as in the raw extract indicating that the anthocyanin structures had not changed during the
loading step. As no anthocyanin polymers were detected, agglomeration of anthocyanins in the
concentrated extract might have been the main cause of the wide shape of the anthocyanin
an
breakthrough curve. Agglomeration depends on the experimental conditions and on the total
amount of anthocyanins and, therefore, this could not be seen on the molar mass analyses.
M
Although the anthocyanin front eluted from the column after a relatively short time, a
considerable amount of anthocyanins could still be adsorbed to the adsorbent. With 10.2 BV/h
ed

flow rate, approximately 45 % of anthocyanins fed to the column were adsorbed in 21 bed
volumes.
pt

Similar breakthrough curves of anthocyanins has been observed also earlier by Wang et al. [22].
Anthocyanin losses due to a wide breakthrough curve were minimized in [22] by recycling the
ce

solution from column outlet to column inlet until no anthocyanins were detected in the outlet
stream. Another option to minimize the anthocyanin losses would be to use a longer column. This
is because the column efficiency, usually quantified as the number of theoretical plates (NTP), is
Ac

proportional to the column length [41] and, therefore, the dynamic loading capacity of the
adsorbent increases with increasing column length.

Glucose, which is the main impurity in anthocyanin extracts of purple-fleshed potato, was not
adsorbed to the XAD-7HP adsorbent (Fig. 6). Thus, a good separation of anthocyanins from
glucose could be obtained. After the loading step, the nonadsorbed solutes were removed from
the adsorbent bed by water wash. The washing caused no desorption of adsorbed anthocyanins
(data not shown).
18

Page 18 of 37
<<Figure 6 around here>>

Desorption of anthocyanins from XAD-7HP packed column was carried out with aqueous
ethanol solutions with different ethanol concentrations and with and without acetic acid (Fig. 7
and Table 2).The effect of acetic acid addition (7 vol.%) on the desorption of anthocyanins was

t
investigated with 75 vol.% ethanol. Acetic acid increased the amount of desorbed anthocyanins

ip
by over 40 % (Fig. 7 and Table 2). This effect most probably resulted from the dependence of the

cr
structure and stability of anthocyanins on pH. At the pH of the acidic ethanol eluent (pH ~ 3.1),
the anthocyanin molecules are mainly in quinoidal base form [36]. Anthocyanins in this form

us
have weaker affinity towards the XAD-7HP adsorbent than at higher pH at which the
anthocyanins are either in carbinol pseudobase form (pH ~ 5) or chalcone form (pH ~ 6) [36].

an
100 % desorption of the adsorbed anthocyanins from XAD-7HP bed could not be obtained with
any of the acidic ethanol eluents (Fig. 7 and Table 2). This was also visually confirmed from the
residual purple color in the adsorbent bed after the desorption experiments. Approximately same
M
amount of anthocyanins could be desorbed with 50 vol.% and 75 vol.% acidic ethanol solutions
(Table 2). Elution of anthocyanins with 50 vol.% acidic ethanol required more eluent in total
ed

(Fig. 7), but slightly lower amount of pure ethanol (2.25 BV) than elution with 75 vol.% acidic
ethanol (2.4 BV). However, with 75 vol.% acidic ethanol, a more concentrated anthocyanin
solution was obtained (Table 2). Efficient desorption could not be achieved with the 25 vol.%
pt

acidic ethanol (Fig. 7 and Table 2): less than 40 % of adsorbed anthocyanins were eluted in
8.5 BV. No glucose was detected in the samples taken from the desorption experiments,
ce

confirming that no glucose was adsorbed to the XAD-7HP adsorbent.

Ethanol concentration higher than 75 vol.% was not used in the desorption experiments as with
Ac

both 50 vol.% and 75 vol.% acidic ethanol approximately the same amount of anthocyanins could
be desorbed (Table 2). In addition, tailing of anthocyanins (Fig. 7) was not seen with neither of
these eluents, on contrast to the 25 vol.% acidic ethanol. This means that ethanol concentration
higher than 75 vol.% would not have improved the desorption considerably.

In conclusion, efficient desorption of anthocyanins of purple-fleshed potato from XAD-7HP

packed column could be effectively carried out using 75 vol.% acidic (7 vol.% acetic acid)
ethanol. This was due to the short elution time and high anthocyanin concentration in the eluent.
19

Page 19 of 37
<<Figure 7 around here>>

<<Table 2 around here>>

The purified anthocyanin solution was concentrated to a final anthocyanin concentration of

t
1750 mg/L using evaporation under vacuum (T = 35 °C). Molar mass analyses showed no

ip
polymerized anthocyanins in the purified solution. In fact, in this solution, 100 % of the mass
fraction was due to molecules with molar mass in range of 400-600 g/mol. Thus, practically all of

cr
the proteins could be separated from the anthocyanins by adsorption. Polydispersity of the small
molecules was approximately the same as in the raw extract (1.4) clearly indicating that no

us
changes in the anthocyanin structures occurred during the purification. Also, according to HPLC-
MS analyses, the main anthocyanins in the purified extract (Fig. 8A) were the same as in the
unpurified extract (see Fig. 4A). an
Glucose concentration in the purified and concentrated extract was 0.05 g/L; thus, over 99.8 % of
M
glucose could be removed. The amount of nitrogen containing species in the purified and
concentrated anthocyanin solution was 459.7 mg/L, which is approximately the same as in the
ethanol-free raw extract. When the concentration factors are taken into account, this means that
ed

over 87 % of the nitrogen-containing species could be removed. Additionally, the unidentified

compounds visible in the HPLC chromatogram of the raw extract (see Fig. 4B) were removed by
pt

adsorption (Fig. 8B).

<<Figure 8 around here>>

ce

Scale up of the extraction and purification process was demonstrated by producing purified
Ac

anthocyanins from 19 kg of purple-fleshed potatoes in a single batch. The mass of potatoes after
freeze-drying was 5.2 kg. The anthocyanin concentration in the ethanol-free raw extract was
560 mg/L; the anthocyanins yield before the purification was 0.38 mg/g fresh potato.

In the adsorption step, with XAD-16 as the adsorbent, similar breakthrough curve was observed
as with XAD-7HP (not shown). Elution of anthocyanins was carried out using 75 vol.% acidic
(7 vol.% acetic acid) ethanol solution. After desorption and evaporation steps, the total
anthocyanin concentration in the purified extract was 4660 mg/L (V = 1.2 L). Thus, the overall

20

Page 20 of 37
yield of anthocyanins was 0.286 mg/g fresh potato and approximately 28 % of anthocyanins were
lost during the purification. The largest loss was caused by the incomplete elution of adsorbed
anthocyanins from the loaded XAD-16 and due to the wide breakthrough curve of anthocyanins
in the loading step. With XAD-7HP, the losses occurring in desorption would have been lower as
the desorption ratio of XAD-7HP was approximately 35 % higher than with XAD-16 (see Fig. 5).

t
ip
Over 99.9 % of glucose and over 85 % of the nitrogen-containing compounds could be removed
with XAD-16. In this respect XAD-16 and XAD-7HP performed similarly.

cr
4 Conclusions

us
Extraction of anthocyanins from purple-fleshed potato (Finnish landrace cultivar Synkeä Sakari)
by solvent extraction and purification of the obtained extracts by adsorption were investigated in
this work. an
The effects of tuber pretreatment method, ethanol concentration, and extraction time on the
M
anthocyanin extraction were investigated. The pretreatment method was found to have the most
significant effect on the extraction yield of anthocyanins. Boiling of potato tubers was found to
be the most efficient pretreatment method.
ed

Sufficient extraction time was found to be six hours. A longer time resulted in only a minor
increase in the anthocyanin concentration. Ethanol concentration was also found to have only a
pt

minor effect on the extraction yield. The extraction could be done efficiently with 20 vol.% acidic
(7 vol.% acetic acid) ethanol.
ce

Six commercially available adsorbents were tested for the adsorption of anthocyanins. Of these,
Ac

acrylic XAD-7HP was found to be the most efficient. In column mode, slow mass transfer and
agglomeration of anthocyanins in raw extract were found to complicate the purification and to
cause anthocyanin losses.

Desorption of adsorbed anthocyanins from XAD-7HP could be carried out with 75 vol. % acidic
(7 vol. % acetic acid) ethanol. However, complete elution could not be achieved. Although a high
ethanol concentration must be used in the desorption step, chemical consumption of the

21

Page 21 of 37
purification step should not be high as ethanol could be separated from the purified solution by
nanofiltration or evaporation and recycled.

Purification of anthocyanins by adsorption removed most of the major impurities from the
anthocyanin solution. This shows that adsorption with XAD-7HP polymeric adsorbent is an

t
efficient method for obtaining pure anthocyanin rich extracts from purple-fleshed potatoes.

ip
The method presented in this study is an efficient way to produce highly valuable acylated

cr
anthocyanins from low value raw material (purple-fleshed potato). This process could also be
applied on a larger scale as the scale up of the process can be done easily.

us
Acknowledgement

an
The authors thank Liz Gutiérrez Quequezana (M.Sc.) for carrying out the HPLC analysis and
HPLC-MS analysis of the anthocyanin extract, and Ms. Heini Issakainen and Ms. Tea Mäkelä for
carrying out the extraction and purification of anthocyanins from the large batch of potatoes.
M
References
ed

1 Valls, J., Millán, S., Pilar Martí, M., Borrás, E., Arola, L., Advanced separation methods
of food anthocyanins, isoflavones and flavanols, J. Chrom. A 1216 (2009), 7143-7172.
2 Silva, E.M., Pompeu, D.R., Larondelle, Y., Rogez, H., Optimisation of the adsorption of
pt

polyphenols from Inga edulis leaves on macroporous resins using an experimental design
methodology, Sep. Purif. Technol. 53 (2007), 274-280.
3 Lapornik, B., Prosek, M., Wondra, A.G., Comparison of extracts prepared from plant by-
ce

products using different solvents and extraction time, J. Food Eng. 71 (2005), 214-222.
4 Fan, G., Han, Y., Gu, Z., Chen, D., Optimizing conditions for anthocyanins extraction
from purple sweet potato using response surface methodology (RSM), LWT-Food Sci.
Technol. 41 (2008), 155-160.
Ac

5 Burgos, G., Amoros, W., Muñoa, L., Sosa, P., Cayhualla, E., Sanchez, C., Díaz, C.,
Bonierbale, M., Total phenolic, total anthocyanin and phenolic acid concentrations and
antioxidant activity of purple-fleshed potatoes as affected by boiling, J. Food Compos.
Anal. 30 (2013), 6-12.
6 Andersen, Ø.M., Jordheim, M., The Anthocyanins, in Flavonoids – Chemistry,
biochemistry and applications (Andersen, Ø.M., Markham, K.R., eds.), CRC Press, Boca
Raton USA, 2006, pp. 471-552.
7 Ghosh, D., Konishi, T., Anthocyanins and anthocyanin-rich extracts: role in diabetes and eye
function, Asia Pac. J. Clin. Nutr. 16 (2007), 200-208.
8 Castañeda-Ovando, A., Pacheco-Hernández, M.D.L., Páez-Hernández, M.E., Rodríguez,
J.A., Galán-Vidal, C.A., Chemical studies of anthocyanins: A review, Food Chem. 113
(2009), 859-871.
22

Page 22 of 37
9 Bridgers, E.N., Chinn, M.S., Truong, V.D., Extraction of anthocyanins from industrial
purple-fleshed sweet potatoes and enzymatic hydrolysis of residues for fermentable sugars,
Ind. Crop. Prod. 32 (2010), 613-620.
10 Puértolas, E., Cregenzán, O., Luengo, E., Álvarez, I., Raso, J., Pulsed-electric-field-
assisted extraction of anthocyanins from purple-fleshed potato, Food Chem. 136 (2013),
1330-1336.
11 Kang, Y.J., Jung, S.W., Lee, S.J., An optimal extraction solvent and purification adsorbent

t
to produce anthocyanins from black rice (Oryza sativa cv. Heugjinjubyeo), Food Sci.

ip
Biotechnol. 23 (2014), 97-106.
12 Chandrasekhar, J.,Madhasudhan, M.C., Raghavarao, K.S.M.S, Extraction of anthocyanins
from red cabbage and purification using adsorption, Food Bioprod. Process. 90 (2012), 615-

cr
623.
13 Patil, G., Madhusudhan, M.C., Ravindra, B., Raghavarao, K.S.M.S, Extraction,

us
dealcoholization and concentration of anthocyanin from red radish, Chem. Eng. Process. 48
(2009), 364-369.
14 Li, J., Li, X.-D., Zhang, Y., Zheng, Z.-D., Qu, Z.-Y., Liu, M., Zhu, S.-H., Liu, S., Wang,
M., Qu, L., Identification and thermal stability of purple-fleshed sweet potato

15
136 (2013), 1429-1434.
an
anthocyanins in aqueous solutions with various pH values and fruit juices, Food Chem.

Todaro, A., Cimino, F., Rapisarda, P., Catalano, A.E., Barbagallo, R.N., Spagna, G.,
Recovery of anthocyanins from eggplant peel, Food Chem. 114 (2009), 434-439.
M
16 Truong, V.D., Hu, Z., Thompson, R.L., Yencho, G.C., Pecota, K.V., Pressurized liquid
extraction and quantification of anthocyanins in purple-fleshed sweet potato genotypes, J.
Food Compos. Anal. 26 (2012), 96-103.
17 Buran, T.J., Sandhu, A.K., Li, Z., Rock, C.R., Yang, W.W., Gu, L., Adsorption/desorption
ed

characteristics and separation of anthocyanins and polyphenols from blueberries using

macroporous adsorbent resins, J. Food Eng. 128 (2014), 167-173.
18 Hillebrand, S., Naumann, H., Kitzinski, N., Köhler, N., Winterhalter, P., Isolation and
characterization of anthocyanins from blue-fleshed potatoes (Solanum tuberosum L.), Food 3
pt

(2009), 96-101.
19 Jampani, C., Naik, A., Raghavarao, K.S.M.S., Purification of anthocyanins from jamun
(Syzygium cumini L.) employing adsorption, Sep. Purif. Technol. 125 (2014), 170-178.
ce

20 Liu, X., Xiao, G., Chen, W., Xu, Y., Wu, J., Quantification and purification of mulberry
anthocyanins with macroporous resins, J. Biomed. Biotechnol. 5 (2004), 326-331.
21 Coutinho, M.R., Quadri, M.B., Moreira, R.F.P.M., Quadri, M.G.N., Partial purification of
Ac

anthocyanins from Brassica oleracea (red cabbage), Separ. Sci. Technol. 39 (2004), 3769-
3782.

23

Page 23 of 37
22 Wang, X., Hansen, C., Allen, K., Extraction of anthocyanins from black bean canning
wastewater with macroporous resins, J. Food Sci. 79 (2014), 184-188.
23 Liu, X., Xu, Z., Gao, Y., Yang, B., Zhao, J., Wang, L., Adsorption characteristics of
anthocyanins from purple-fleshed potato (Solanum tuberosum Jasim) extract on macroporous
resins, Int. J. Food Eng. 3 (2007), 1556-3758.
24 Lewis, C. E., Walker, J. R. L., Lancaster, J. E., Sutton, K. H., Determination of

t
anthocyanins, flavonoids and phenolic acids in potatoes. Ⅰ. Coloured cultivars of

ip
Solanum tuberosum L. J. Agr.Food Chem. 77 (1998), 45-57.
25 Giusti, M.M., Wrolstad, R.E., Acylated anthocyanins from edible sources and their

cr
application in food systems, Biochem. Eng. J. 14 (2003), 217-225.
26 Lee, J., Durst, R.W., Wrolstad, R.E., Determination of total monomeric anthocyanin
pigment content of fruit juices, beverages, natural colorants, and wines by the pH

us
differential method: collaborative study, J. AOAC Int. 88 (2005), 1269-1278.
27 Hosseinian, F.S., Li, W., Beta, T., Measurement of anthocyanins and other
phytochemicals in purple wheat, Food. Chem. 109 (2008), 916-924.
28 Lisińska, G., Leszczyński, W., Potato Science and technology, Elsevier Science

29
publishers, Essex, United Kingdom, 1989. an
Pal, S., Bhattacharya, A., Konar, A., Mazumdar, D., Das, A.K., Chemical composition of
potato at harvest and after cold storage, Int. J. Vege. Sci. 14 (2008), 162-176.
30 Gooding, E.G.B., MacDougall, D.B., The accelerated freeze-drying of potatoes, Eur.
M
Potato J. 4 (1961), 69-73.
31 Brás, T., Fernandes, M.C., Santos, J.L.C., Neves, L.A., Recovering bioethanol from olive
bagasse fermentation by nanofiltration, Desalination and Water Treatment 51 (2013),
ed

4333-4342.
32 Yanagida, A., Shoji, T., Kanda, T., Characterization of polymerized polyphenols by size-
exclusion HPLC, Biosci. Biotechnol. Biochem. 66 (2002), 1972-1975.
33 Tsanova-Savova, S., Dimov, S., Ribanova, F., Anthocyanins and color variables of
pt

Bulgarian aged red wines, J. Food Compos. Anal. 15 (2002), 647-654.

34 Tsai, P.-J., Huang, H.-P., Huang, T.-C., Relationship between anthocyanin patterns and
antioxidant capacity in mulberry wine during storage, J. Food Quality 27 (2004), 497-
ce

505.
35 He, F., Liang, N.-N., Mu, L., Pan, Q.-H., Wang, J., Reeves, M.J., Duan, C.-Q.,
Anthocyanins and their variation in red wines II. Anthocyanin derived pigments and their
color evolution, Molecules 17 (2012), 1483-1519.
Ac

36 Moldovan, B., David, L., Chisbora, C., Cimpoiu, C., Degradation kinetics of anthocyanins
from European cranberrybush (Viburnum opulus L.) fruit extracts. Effects of temperature, pH
and storage solvent, Molecules 17 (2012), 11655-11666.
37 The Dortmund Data Bank, Dortmund Data Bank Software & Separation Technology,
http://www.ddbst.com/ddbst.html, accessed 22.3.2016.
38 Partras, A., Brunton, N.P., O’Donnell, C., Tiwari, B.K., Effect of thermal processing on
anthocyanin stability in foods; mechanisms and kinetics of degradation, Trends Food Sci.
Tech. 21 (2010), 3-11.
39 Cao, S.-q., Liu, L., Pan, S.-y., Thermal degradation kinetics of anthocyanins and visual
color of blood orange juice, Agr. Sci. China 10 (2011), 1992-1997.

24

Page 24 of 37
40 Fischer, U.A., Carle, R., Kammerer, D.R., Thermal stability of anthocyanins and
colourless phenolics in pomegranate (Punica granatum L.) juices and model solutions,
Food Chem. 138 (2013), 1800-1809.
41 Guiochon, G., Felinger, A., Shirazi, D.G., Katti, A.M., Fundamentals of preparative and
nonlinear chromatography, 2nd ed., Academic Press, Amsterdam, Netherlands, 2006.

t
ip
cr
us
an
M
ed
pt
ce
Ac

25

Page 25 of 37
FIGURE CAPTIONS

Figure 1. Effect of potato pretreatment method on the amount of anthocyanins extracted

t
from the purple-fleshed potato Synkeä Sakari. Experimental conditions:

ip
completely pigmented potatoes; ethanol:water:acetic acid ratio =
40 vol.%: 53 vol.%: 7 vol.%; solid–liquid ratio = 1:5 (wt.:wt.); room temperature;
extraction time = 6 hours. The error bars mark the standard error.

cr
Figure 2. Effect of time on the amount of anthocyanins extracted from the purple-fleshed

us
potato Synkeä Sakari. Experimental conditions: completely pigmented potato;
pretreatment method = boiling; ethanol:water:acetic acid ratio =
40 vol.%: 53 vol.%: 7 vol.%; solid–liquid ratio = 1:5 (wt.:wt.); room temperature.
The error bars mark the standard error.
an
Figure 3. Effect of ethanol concentration on the amount of anthocyanins extracted from the
purple-fleshed potato Synkeä Sakari. Experimental conditions: pretreatment
M
method = boiling; 7 vol.% acetic acid in the extraction medium; solid–liquid ratio
= 1:5 (wt.:wt.); room temperature; extraction time = 24 hours. The error bars mark
the standard error.
ed

Figure 4. Liquid chromatography analysis of the raw extract obtained from purple-fleshed
potatoes. A: HPLC chromatogram on Kinetex C18 column; peaks identified using
pt

MS; UV signal. Numbered peaks: 1 = petunidin-coumaroylrutinoside-glucoside;

2 = peonidin-p-coumaroylrutinoside-glucoside. B: HPLC chromatogram on
Shodex SP-0810 column; RI signal. Numbered peaks: 1 = unidentified impurity;
ce

2 = glucose; 3 = unidentified impurity; 4 = ethanol.

Figure 5. Adsorption of anthocyanins from purple-fleshed potato extract (black) and

Ac

desorption of the adsorbed anthocyanins (grey) with various adsorbents.

Experimental conditions in adsorption: batch equilibrium experiments;
Canthocyanins,initial = 775.2 mg/L; T = 23 °C; madsorbent = 4 g; msolution = 16 g;
teq = 15 hours. Experimental conditions in desorption with 50 vol.% acidic
(7 vol.% acetic acid) ethanol: batch equilibrium experiments; T = 23 °C; madsorbent
= 4 g; methanol = 14.6 g; teq = 1 hour. See Table 1 for adsorbent details. The error
bars mark the standard error.

Figure 6. Breakthrough curves of anthocyanins and glucose (open grey triangles; with
10.2 BV/h flow rate) from a column packed with XAD-7HP adsorbent with
1.7 BV/h (open black square), 6.8 BV/h (grey diamond), and 10.2 BV/h (black
26

Page 26 of 37
 circles) flow rates. Feed concentrations: 224.5 mg/L anthocyanins, 3.0 g/L
glucose. Experimental conditions: T = 23 °C; top–down flow. C/Cfeed = outlet
concentration relative to feed concentration. Average analytical error in the
anthocyanin and glucose concentrations: 5.8 % and 0.6 %, respectively.

Figure 7. Desorption of the adsorbed anthocyanins from XAD-7HP adsorbent in a column.

t
Initial amount of anthocyanins in the adsorbent = 18.5 mg. Experimental

ip
conditions: T = 23 °C; flow rate = 3.4 BV/h; top–down flow. Symbols: open grey
diamond = 25 vol.% acidic ethanol; grey triangle = 50 vol. % acidic (7 vol.%

cr
acetic acid) ethanol; filled black circle = 75 vol.% acidic ethanol; open black
circle = 75 vol.% non-acidic ethanol. C/Cfeed and average analytical error in the
anthocyanin and glucose concentrations: see caption of Fig. 6.

us
Figure 8. Liquid chromatography analysis of the purple-fleshed potato extract purified by
adsorption. A: HPLC-MS chromatogram on Kinetex C18 column; peaks identified
an
using MS; UV signal. Numbered peaks: 1 = petunidin-coumaroylrutinoside-
glucoside; 2 = peonidin-p-coumaroylrutinoside-glucoside.
chromatogram on Shodex SP-0810 column; RI signal. Numbered peaks:
B: HPLC

1 = glucose; 2 = ethanol.
M
ed
pt
ce
Ac

27

Page 27 of 37
TABLES

Table 1. Main properties of the adsorbents used in this work. Data

provided by the suppliers.
Trade name Supplier Matrix Functional Surface Pore size,

t
2
groups area, m /g Å

ip
XAD-4 Sigma- Polystyrene– - ≥750 100
Aldrich divinylbenzene

cr
XAD-7HP Sigma- Acrylic ester - ≥380 300-400
Aldrich
XAD-16N Sigma- Polystyrene– - ≥800 200

us
Aldrich divinylbenzene
MN-200 Purolite Polystyrene– - ≥900 800*/15**
divinylbenzene
Amberlyst 15 Sigma- Polystyrene– Sulfonic 53 300
(H+ form)
CS16GC
Aldrich
Finex
divinylbenzene
Polytyrene–
an acid
Sulfonic n.a. Gel type
(Na+ form) divinylbenzene acid
*
M
Meso- and macropores
**
Micropores
ed
pt
ce
Ac

28

Page 28 of 37
Table 2. Effect of ethanol and acetic acid concentrations on the desorption of
anthocyanins from XAD-7HP adsorbent.
C, vol. % Canthocyanins, mg/L
Recovery, %
EtOH AcOH Maximum Average
25 7 35.2 73.7 -

t
50 7 71.4 656.5 176.1

ip
75 7 72.2 788.3 240.7
75 0 50.0 568.8 228.5

cr
us
an
M
ed
pt
ce
Ac

29

Page 29 of 37
Figure1

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 30 of 37
Figure2

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 31 of 37
Figure3

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 32 of 37
Figure4
mAU

us
6
A 1

5
2

an
4

M
2

d
10 15 20 25 30 35 40 45 50 55 60 65 min

nRIU e
pt
4
7e4 B
ce

6e4

5e5
Ac

4e4

3
3e4
1
2e4

1e4 2

0
Page 33 of 37
0 5 10 15 20 25 30 min
Figure5

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 34 of 37
Figure6

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 35 of 37
Figure7

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 36 of 37
Figure8
mAU

us
4
A 1
2

an
3

M
1

d
0 10 20 30 40 50 60 70 min

nRIU e
pt
6e5 2
B
ce

5e5

4e5
Ac

3e5

2e5

1e5

1
0
Page 37 of 37
0 5 10 15 20 25 30 min