Food Chemistry 192 (2016) 171177

Contents lists available at ScienceDirect

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

Biotransformation of anthocyanins from two purple-eshed sweet

potato accessions in a dynamic gastrointestinal system
Stan Kubow a,, Michle M. Iskandar a, Kebba Sabally a, Behnam Azadi a, Shima Sadeghi Ekbatan a,
Premkumari Kumarathasan b, Dharani Dhar Das b, Satya Prakash c, Gabriela Burgos d, Thomas zum Felde d
a

School of Dietetics and Human Nutrition, McGill University, 21111 Lakeshore, Ste. Anne de Bellevue, QC H9X 3V9, Canada
Analytical Biochemistry and Proteomics Laboratory, Mechanistic Studies Division, Healthy Environments and Consumer Safety Branch, Health Canada, 50 Colombine Drwy,
Ottawa, ON K1A 0K9, Canada
c
BioMedical Engineering Department, McGill University, 3775 University Street, Room 311, Montreal, QC H3A 2B4, Canada
d
International Potato Center (CIP), Avenida La Molina 1895, La Molina, Apartado Postal 1558 Lima, Peru
b

a r t i c l e

i n f o

Article history:
Received 1 April 2015
Received in revised form 26 June 2015
Accepted 29 June 2015
Available online 30 June 2015
Keywords:
Ipomoea batatas L.
Purple eshed sweet potato
Anthocyanins
Bioaccessibility
Biotransformation
Human gastro-intestinal model

a b s t r a c t
Cooked, milled purple-eshed sweet potato (PFSP) accessions, PM09.812 and PM09.960, underwent
digestion in a dynamic human gastrointestinal (GI) model that simulates gut digestive conditions to
study the bioaccessibility and biotransformation of anthocyanins. Matrix-assisted laser desorption
ionization time-of-ight mass spectrometry showed accession-dependent variations in anthocyanin
release and degradation. After 24 h, more anthocyanin species were detected in the small intestinal vessel
relative to other vessels for accession PM09.960 whereas more species appeared in the ascending colonic
vessel for accession PM09.812. The ferric reducing antioxidant power was increased in the small
intestinal vessel for PM09.960 and in the ascending colonic vessel for accession PM09.812, corresponding
to the appearance of a majority of anthocyanins for each accession. These results show that intestinal
and colonic microbial digestion of PFSP leads to an accession-dependent pattern for anthocyanin
bioaccessibility and degradation.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Purple eshed sweet potatoes (PFSP) contain a signicant content of anthocyanins comparable to that of other high anthocyanin
containing fruits and vegetables such as grapes, plums, sweet cherries, raspberries and eggplant (Truong et al., 2010). Anthocyanins
from PFSP have been shown to exhibit stronger radicalscavenging activity than anthocyanin pigments from red cabbage,
elderberry, grape skin and purple corn (Kano, Takayanagi, &
Harada, 2005).
Despite their high in vitro antioxidant properties, there is limited research regarding how digestive processes of PFSP affect their
antioxidant activity and anthocyanin structures, which could be
altered during gastrointestinal (GI) digestion. During GI digestive
processes involving pH changes, digestive enzymes and microbial
Corresponding author.
E-mail addresses: stan.kubow@mcgill.ca (S. Kubow), michele.iskandar@mail.
mcgill.ca (M.M. Iskandar), kebba.sabally@mcgill.ca (K. Sabally), behnam.azadi@
mcgill.ca (B. Azadi), shima.sadeghi@mail.mcgill.ca (S. Sadeghi Ekbatan),
premkumari.kumarathasan@hc-sc.gc.ca (P. Kumarathasan), dharani.das@hc-sc.gc.
ca (D.D. Das), satya.prakash@mcgill.ca (S. Prakash), g.burgos@cgiar.org (G. Burgos),
t.zumfelde@cgiar.org (T. zum Felde).
http://dx.doi.org/10.1016/j.foodchem.2015.06.105
0308-8146/ 2015 Elsevier Ltd. All rights reserved.

metabolism, anthocyanins are released from the food matrix and

undergo extensive degradation that can affect their bioactivities.
Furthermore, anthocyanin metabolites could also be bioavailable
for absorption and exert protective antioxidant and
anti-inammatory effects on intestinal disorders such as inammatory bowel disease (Wu, Xu, Dong, He, & Yu, 2011) and colon
cancer (Lim et al., 2013).
In vitro GI models used to evaluate digestion of polyphenols
such as anthocyanins have utilized dual enzyme pepsinpancreatin digestion to simulate gastric and small intestinal tract digestion that could also be coupled with batch colonic reactors with
human gut microora to assess the impact of human microbial
metabolism (Alminger et al., 2014; Tarko, Duda-Chodak, & Zajac,
2013). Such studies have shown major variations in anthocyanin
stability of different foods during digestive processes, which
appear to be highly dependent on food matrix composition and
structure. For example, anthocyanin-rich extracts of red cabbage
showed extensive degradation under the mild alkaline conditions
of pancreatic digestion whereas most anthocyanins of raw red
cabbage reached colonic vessels intact to be extensively degraded
by bacteria (Pods?dek, Redzynia, Klewicka, & Koziokiewicz,
2014).

172

S. Kubow et al. / Food Chemistry 192 (2016) 171177

To date, studies of anthocyanin digestion have not utilized

computer-controlled dynamic multistage continuous digestion
models that can better simulate in vivo conditions (Alminger
et al., 2014). The dynamic in vitro digestion system mimics the
in vivo dynamics of transit during digestion and considers the varying microbial or digestive conditions in different segments of the
GI tract including the ascending colon, transverse colon and
descending colon. Adjustment of the pertinent pH for each colonic
region leads to differences in growth of corresponding microbial
communities that results in alterations in the metabolic activity
of the colonic bioreactors (Blanquet-Diot et al., 2009; Martoni
et al., 2007; Molly, Vande Woestyne, De Smet, & Verstraete, 1994).
In the present study, a computer controlled dynamic human GI
model was used to evaluate the effect of digestion on the release
and biotransformation of anthocyanins present in cooked, milled
and freeze-dried samples of two PFSP accessions (PM09.812 and
PM09.960) conserved in the in-trust germplasm collection at
International Potato Center (CIP). Following digestion of the sweet
potato accessions in the GI model, samples from all the compartments of the GI model (stomach, small intestine, ascending, transverse and descending colon) were assessed for anthocyanin
proles using matrix assisted laser desorption ionization time of
ight mass spectrometry (MALDI-TOFTOF-MS) and measured
for their antioxidant capacity via the ferric reducing antioxidant
power (FRAP) assay.

PM09.812

2. Materials and methods

2.1. Plant material
Two PFSP accessions: PM09.812 (CIP number: 109264.3) and
PM09.960 (CIP number: 109270.2) were used for this study
(Fig. 1). The two CIP breeding clones with the same mother but different fathers were selected for their good yield and deep purple
colors (PM09.812: female (SM07.048, CIP107791.1)  male
(SM07.524; CIP107805.1); PM09.960: female (SM07.048;
CIP107791.1)  male (SM07.626; CIP107809.3)). The two accessions were grown in San Ramon, Junin, Peru and harvested in
October 2013. Ten roots per accession were collected and brought
to the Quality and Nutrition Laboratory of CIP in Lima, Peru where
they were cooked. Unpeeled roots of each accession were placed in
stainless steel pots containing boiling water. One control for each
accession was used to determine the cooking time required.
Roots were considered as cooked when a stainless steel stick could
penetrate the control roots easily. Following cooking, the roots
were peeled, representatively sampled, freeze dried and milled
through 40 mesh. Freeze dried and milled samples of each accession were sent to McGill University.
2.2. Computer controlled dynamic human gastrointestinal model
The components of the simulated human GI model involved ve
reactors, the last three of which contain microbiota of a different
part of the human GI tract. In order of sequence (reactors 15):
the stomach (V1), small intestine (V2; duodenum, jejunum and
ileum), the ascending (V3), the transverse (V4) and the descending
colon (V5). The system is fully computer-controlled and has been
validated using enumeration procedures, short chain fatty acid
production patterns, enzymatic activities, gas production, and by
microorganism-associated activities (Blanquet-Diot et al., 2009;
Martoni et al., 2007; Molly et al., 1994).
Five healthy, non-smoking individuals with no history of GI disease or antibiotic use in the previous 6 months provided fecal samples for the study. Fecal samples were freshly collected, pooled,
and readily used to prepare the fecal solution. The system

PM09.960
Fig. 1. The purple-eshed sweet potato accessions PM09.812 and PM09.960.

underwent a 2-week stabilization period to allow bacterial communities from the diluted human fecal samples to grow and
stabilize.
During and following stabilization, the microbial ecosystem
was sustained by feeding the system with 300 mL of sterile medium (set at pH 2 before autoclaving and stored at 37 C) every
8 h. The composition of the GI nutrient solution essential for bacterial survival is shown in Table 1.
The fermentation vessels were maintained anaerobic by purging the headspace with oxygen-free nitrogen and stirring continuously on magnetic stirrers. The temperature of the simulator was
kept at 37 C. Upon entering vessel 2 (small intestine), pancreatic
juice supplemented with bile (12 g/L NaHCO3, 0.9 g/L pancreatin;
SigmaAldrich) and 6 g/L Oxgall (Difco) were added to neutralize

Table 1
Composition of the GI nutrient solution essential for bacterial survival.
Component

Concentration (g/L)

Arabinogalactan
Pectin
Starch
Xylan
Glucose
Yeast extracts
Peptone
Mucin
Cysteine

1
2
3
1
0.4
3
1
4
0.5

Source: Molly et al. (1994).

173

S. Kubow et al. / Food Chemistry 192 (2016) 171177

stomach acidity. In this way, the pH of the rst two vessels was set
by the input of either supply medium or pancreatic juice. Vessels 3,
4 and 5 (representing the colon) were pH-controlled between 5.6
and 5.9; 6.1 and 6.4; and 6.6 and 6.9, respectively. The pH was
measured with a probe connected to a pH meter and was automatically adjusted by adding 0.2 M NaOH or 0.5 M HCl.
The cooked, freeze-dried, and milled samples of each accession
(18.15 g) were digested in the GI model. The amount provided to
the GI model was based on the fact that only a small amount of
PFSP is included in the human diet. Considering that the accessions
used in this study have around 35% dry matter, the amount provided to the GI model represent approximately 50 g of each accession on a fresh weight basis.
On the day of treatment the freeze-dried sweet potato samples
were incorporated into the GI food solution and subjected to 24 h
digestion by the GI model. Samples were collected from all the
compartments of the GI model before addition of the
freeze-dried potato (T = 0 control) and after 8 h (T = 8) and 24 h
(T = 24) of digestion of the freeze-dried potato. One day of treatment was followed by a 3-day washout period, during which the
system was fed the control GI nutrient solution without potato.
Each treatment lasted 24 h and sampling was performed at 0, 8
and 24 h throughout the day of treatment (Fig. 2).
For each day of treatment, an aliquot (20 mL) was removed
from vessels 1 through 5, and stored at 80 C for later analysis
and characterization of polyphenolic content. To prevent photodecomposition of the polyphenols, all of the digestive compartments
and collection vessels were wrapped in tin foil.
2.3. Fecal water (FW) preparation
Fecal water (FW) was prepared by collecting samples of the
contents of vessels 1 through 5 and centrifuging them at 200g for
20 min; the supernatants were stored at 80 C until use for
MALDI-TOFTOF-MS and antioxidant capacity analyses.

formic acid in methanol. The eluted samples were dried under

nitrogen and stored at 80 C until MALDI-TOFTOF-MS analysis.
2.5. MALDI-TOFTOF-MS analysis
MALDI-TOFTOF-MS analysis was performed based on the
method of Kumarathasan et al. (2014) with modications. Dried
samples were re-solubilized in 50% acetonitrile/0.1% triuoroacetic
acid (TFA)/H2O prior to MALDI-TOFTOF-MS analysis, 1 lL of sample was spotted in duplicates at the center of the sample location
on a 384/600 Anchor Chip target plate (Bruker Daltonics). One lL
of matrix solution (0.5 mg/mL a-cyano-4-hydroxycinnamic acid
in 50% acetonitrile/0.1% TFA/H2O) was added on top of the sample
spot and was mixed by pulling and releasing the liquid with the
help of the pipette tip. The sample spots were allowed to dry by
evaporation in open air. An on-target washing of the sample spots
was carried out by placing 2.5 lL of cold 1% TFA in water on the
dried sample spot, and removing the liquid after 10 s. Washed
spots were dried in open air and the mass spectral proles were
recorded using Autoex III time-of-ight mass spectrometer
(Bruker Daltonics, Bremen, Germany) equipped with a Smart
Beam laser emitting at 355 nm wavelength, a 1 GHz sampling
rate digitizer, a pulsed ion extraction source, and a TOFTOF-MS
analyzer. The instrument calibration was done using external calibration standards (Bruker Daltonics). Detection was carried out in
a reectron positive mode. In a typical experiment, a composite
spectrum (total of 1000 shots) was obtained by summation of ve
200-shots of individual spectra. The sampling sites were selected
randomly for every sample in order to obtain homogenous acquisition. The acquisition of mass spectra was carried out using the Flex
Control 3.3 software whereas further processing was done using
Flex Analysis 3.3 software. Anthocyanins in the samples were identied based on their masses and information obtained from literature. The anthocyanins were quantied using the external
standard, keracyanin chloride (cyanidin-3-O-rutinoside chloride)
(KCC-Eq).

2.4. Solid phase extraction procedure for the FW samples

2.6. Ferric reducing ability of plasma (FRAP) assay
Solid phase extraction was performed based on the method of
Mullen, Edwards, Serani, and Crozier (2008) with modications.
Equal amounts of FW and phosphate buffer pH 6 (5 mL:5 mL) were
applied to a 100 mg/3 mL of C18-E cartridges, which were conditioned with 300 lL of 5% formic acid and equilibrated with:
300 lL of 5% formic acid in methanol. The cartridges were washed
twice with 300 lL of 5% formic acid and dried for 3 min. Bound
anthocyanins were then eluted using two folds of 300 lL of 5%

The ferric reducing ability of plasma (FRAP) antioxidant capacity assay was used to determine the total antioxidant potential in
the supernatant of the fecal water obtained from the gut model.
Briey, the electron-donating capacity of an antioxidant is measured by the change in absorbance at 593 nm when a
blue-colored Fe2+-tripyridyltriazine (Fe2+TPTZ) compound was
formed from a colorless oxidized Fe3+ form (Benzie & Strain,

2 weeks:

Time 0h:

Time 8h:

Stabilization
period:

Collection of
control
samples from
all vessels

Collection of
samples from
all vessels

Injection of
GI food every
8 hours

Followed by
injection of
GI food
containing
potato meal

Followed by
second
injection of
GI food
containing
potato meal

Time 16h:

Time 24h:

Third
injection of
GI food
containing
potato meal

Collection of
samples from
all vessels

Fig. 2. Schematic representation of gut model experiments timeline and sample collection.

174

S. Kubow et al. / Food Chemistry 192 (2016) 171177

1996). A standard curve was prepared using solutions of ferrous

sulfate at concentrations ranging from 0.1 to 10 mM. The reagent
was prepared with 10:1:1 ratio of 300 mM acetate buffer pH 3.6:
10 mM TPTZ (2,4,6-tripyridyl-s-triazine) solution: 40 mM HCl at
50 C, and 20 mM FeCl36H2O solution. Once the FRAP working
solution was prepared it was immediately incubated for 10 min
at 37 C. Thirty lL H2O, 10 lL standards or samples, and 200 lL
FRAP working solution were added into a 96-well plate. After
30 min incubation absorbance was read at 593 nm in a microplate
reader (Innite PRO 200 series, Tecan Group).

3. Results and discussion

3.1. Anthocyanin proles in digested PM09.812 and PM09.960 sweet
potatoes
As many anthocyanin compounds are not generally commercially available as standards for identication and since the identity of individual anthocyanins using UVvisible spectra is not
denitive, identication of individual anthocyanins was based on
MALDI-TOFTOF-MS data using information obtained from literature pertaining to sweet potato anthocyanins. Previous studies
have shown that the principal anthocyanin structure of sweet
potatoes is anthocyanidin 3-acyl-sophoroside-5-glucoside, with
the acyl substitute being primarily caffeoyl, hydroxybenzoyl, or
feruloyl residues (Montilla, Hillebrand, & Winterhalter, 2011),
which corresponds to the primary anthocyanin structures
observed in the gut model vessels of the sweet potato samples
(Table 2). The anthocyanin proles in the stomach have been well
demonstrated to remain stable after 2 h under simulated gastric
conditions (McDougall, Dobson, Smith, Blake, & Stewart, 2005),
which is likely due to the stability of the avylium cation form of
anthocyanin at low pH (Clifford, 2000).
In the stomach vessel (V1) after 24 h, the major anthocyanins
observed in both accessions have a general anthocyanidin

3-acyl-rutinoside-5-glucoside structure that differs compositionally between the two accessions (Table 2). There were two major
anthocyanidins in PM09.812 in the form of peonidin and
pelargonidin with p-coumaroyl and feruloyl as the acyl residue
whereas peonidin was the primary anthocyanidin present in the
sweet potato accession PM09.960, which contained a feruloyl acyl
substitute. A total of eight acylated anthocyanins were identied in
V1 of the pigmented sweet potato PM09.812 with petunidin,
peonidin, pelargonidin, or cyanidin bases (Table 2). Non- and
mono-acylated petunidin species and mono-acylated peonidin
species accounted for 35% and 36% of total anthocyanins, respectively. Mono-acylated cyanidin and pelargonidin species accounted
for 16% and 13% of total anthocyanins, respectively. The sweet
potato PM09.960 in V1 contained seven acylated anthocyanins
on pelargonidin, peonidin or cyanidin bases. Mono- and
di-acylated peonidin species accounted for 64% of total anthocyanins. Mono- and di-acylated cyanidin species and monoacylated pelargonidin species accounted for 17% and 19% of total
anthocyanins, respectively. The above ndings coincide with previous literature, which indicates that peonidin amounts are generally
higher in PFSP than cyanidin-based anthocyanins (Lee, Park, Choi,
& Jung, 2013).
Peonidin 3-(600 -p-coumaroyl-sophoroside)-5-glucoside and
petunidin-3-p-coumaroyl-rutinoside-5-glucoside with m/z values
of 771 and 933, respectively, were the major anthocyanins seen
in V1 of the sweet potato accession PM09.812, accounting for
54% of the total anthocyanin content in V1. The sweet potato accession PM09.812 also contained lesser quantities of the petunidin
anthocyanidin in V1 in the form of petunidin-3-rutinoside-5-gluco
side (m/z 787) and the peonidin anthocyanidin in the form of
peonidin 3-p-coumaroyl-rutinoside-5-glucoside. Cyanidin anthocyanins were seen in relatively lower amounts in the form of
cyanidin
3-p-hydroxy-benzoyl-sophoroside-5-glucoside
(m/z
893). Smaller quantities of pelargonidin derivatives were also
observed in V1 including pelargonidin-procyanidin-rutinoside
and pelargonidin-feruloyl-rutinoside (m/z 933).

Table 2
Proposed identication of anthocyanin peaks in pigmented sweet potatoes and their concentration in the samples exposed to human simulated intestinal digestion (mg/L).a
(m/z)b

725
731
741
745
755
757
771
773
773
787
801
887
893
907
917
933
935
963
1055
1065
1069
1095
1127
a
b
c
d
e

Proposed compoundc

Pelargonidin-procyanidin-rutinoside
Cyanidin 3-p-hydroxy-benzoyl-sophoroside-5-glucoside
Fragment of pelargonidin 3-feruloyl-sophoroside-5-glucoside
Peonidin 3-p-hydroxy-benzoyl-sophoroside-5-glucoside
Pelargonidin-feruloyl-rutinoside
Pelargonidin 3-sophoroside-5-glucoside
Peonidin 3-(600 -p-coumaroyl-sophoroside)-5-glucoside
Cyanidin 3-sophoroside-5-glucoside
Cyanidin 3-(600 -caffeoyl-sophoroside)-5-glucoside
Petunidin-3-rutinoside-5-glucoside
Fragment of Peonidin 3-(600 -feruloyl-sophoroside)-5-glucoside
Pelargonidin-3-(p-coumaroyl)-rutinoside-5-glucoside
Cyanidin 3-p-hydroxy-benzoyl-sophoroside-5-glucoside
Peonidin 3-p-hydroxy-benzoyl-sophoroside-5-glucoside
Peonidin-3-p-coumaroyl-rutinoside-5-glucoside
Pelargonidin 3-feruloyl-sophoroside-5-glucoside
Cyanidin 3-caffeoyl-sophoroside-5-glucoside
Peonidin 3-(600 -feruloyl-sophoroside)-5-glucoside
Cyanidin-3-(600 -caffeoyl-600 -p-hydroxy-benzoyl-sophoroside)-5-glucoside
Cyanidin 3-(600 ,600 -dicoumaryl-sophoroside)-5-glucoside
Peonidin 3-caffeoyl-p-hydroxybenzoyl-sophoroside-5-glucoside
Cyanidin 3-feruloyl-p-coumaryl-sophoroside-5-glucoside
Peonidin 3-feruloyl-p-caffeoyl-sophoroside-5-glucoside

Sweet potato A: PM09.812

Sweet potato B: PM09.960

V1d

V2

V3

V4

V5

V1

V2

V3

V4

V5

34
31
44
35
42
40

43
38

31
32
38
45
31
38

32

59
40

32
33

30

37

26
28
23
27
24

24
23
32

26
25
25
26
25

27

21

42

32
34

24

27

148

35
33
33
63
41
40

35
50
50
36
39

63
43

55
28

26

27
54
33
33

41
49
27

31

26

25
62
29
85
69
36
27

32
26
35
25

21
21
24
21
23
23

23
23

22
26

55

179

64

75

40
154

Expressed as cyanidin 3-glucoside equivalents.

Determined by MALDI-TOFTOF-MS analysis.
Identication based on previous literature data (Kim et al., 2012; Montilla et al., 2011; Mori et al., 2010; Tian et al., 2005).
V1 = Stomach; V2 = Small intestine; V3 = Ascending colon; V4 = Transverse colon; V5 = Descending colon.
KCC-Eq values <30 represent trace quantities.

S. Kubow et al. / Food Chemistry 192 (2016) 171177

Peonidin 3-feruloyl-p-caffeoylsophoroside-5-glucoside (m/z 1127)

was the primary anthocyanin noted in V1 from the sweet potato
accession PM09.960 that contributed to 54% of the total anthocyanin
content. The sweet potato accession PM09.960 also contained lesser
quantities of the pelargonidin anthocyanidin in V1 in the form of p
elargonidin-feruloyl-rutinoside (m/z 755) and a trace amount of
pelargonidin in the form of pelargonidin 3-feruloyl-sophorosi
de-5-glucoside (m/z 741). Smaller amounts of cyanidin (cyanidin
3-p-hydroxy-benzoyl-sophoroside-5-glucoside (m/z 893)) and
cyanidin-3-(600 -caffeoyl-600 -p-hydroxy-benzoyl-sophoroside)-5glucoside (m/z 1055) and peonidin compounds peonidin 3-phydroxy-benzoyl-sophoroside-5-glucoside (m/z 907) and peonidin
3-caffeoyl-p-hydroxy-benzoyl-sophoroside-5-glucoside (m/z 1069)
were also observed. Cyanidin-3-(600 -caffeoyl-600 -p-hydroxy-benzoy
l-sophoroside)-5-glucoside (m/z 1055) and its deacylated form
cyanidin-3-p-hydroxy-benzoyl-sophoroside-5-glucoside (m/z 893)
were observed in the stomach.
The two major anthocyanins of PM09.812 observed in V1,
peonidin 3-(600 -p-coumaroyl-sophoroside)-5-glucoside (m/z 771)
and petunidin-3-p-coumaroyl-rutinoside-5-glucoside (m/z 933),
appeared to be completely broken down during simulated human
intestinal digestion in the small intestine vessel (V2). In contrast,
for PM09.960, both of the above anthocyanins only rst appeared
in the ascending colon vessel (V3), which could signify that
microbial metabolism allowed for the release of these anthocyanins from the sweet potato food matrix of PM09.960. In V2 of
PM09.812, pelargonidin-3-feruloyl-sophoroside-5-glucoside was
digested to form the fragment of m/z 741. Pelargonidin-3(p-coumaroyl)-rutinoside-5-glucoside (m/z 887), also known as
pelanin, was present only in the V3 vessel for PM09.812 and V2
and V4 vessels for PM09.960.
For both sweet potato accessions, pelargonidin-feruloyl-rutino
side with a mass of 755 was present in signicant amounts in all
vessels (V1V5). The concentration of this anthocyanin showed a
trend to decrease from V1 to V5 for PM09.812, but showed a different pattern in PM09.960 with a tendency to increase in V3 and V4
and a subsequent decrease in V5. The anthocyanin pelargoni
din-procyanidin-rutinoside with a m/z of 725, was found in significant amounts in all digestion vessels from the digestion of
PM09.812 but was only present in V2 and V4 vessels for
PM09.960. The biphasic phenomenon of a drop and a subsequent
increase in concentrations of various anthocyanins seen during
simulated human gut digestion in the present work has been noted
previously (Aura et al., 2005). Such ndings could be related to
enzymatic release from the food matrix, de-conjugation and subsequent re-conjugation of the anthocyanins, microbial anthocyanin
degradation during digestive enzyme and microbial metabolism
and anthocyanin release and re-incorporation into the fecal matrix
that has a high binding capacity for anthocyanins and their
metabolites. Alternatively or in addition, the lack of detection of
certain metabolites at trace levels (KCC-Eq values <30) could be
due to variations at the limits of detection.
Cyanidin 3-p-hydroxy-benzoyl-sophoroside-5-glucoside (m/z
893) was detected only in V1 for both sweet potato accessions
and was present at a particularly high concentration of 75 lg/mL
KCC-Eq for PM09.812. This anthocyanin has been measured in both
uncooked (Montilla et al., 2010) and cooked purple sweet potatoes
(Truong et al., 2010) and was the most thermally stable anthocyanin associated with protection against a colorectal cancer
model from intake of PFSP (Lim et al., 2013). This latter anthocyanin can be digested via deglycosylation to form cyanidin
3-p-hydroxy-benzoyl-sophoroside (m/z 731) (M (893)-162 for the
loss of glucose) (Tian, Konczak, & Schwartz, 2005), which was
detected in all ve sections of the gut model with PM09.812 and
in the V2V5 vessels of PM09.960.

175

Cyanidin 3-sophoroside-5-glucoside with a molecular ion m/z

value of 773 appeared only in V2 of PM09.960. Cyanidin 3-(600 caff
eoyl-sophoroside)-5-glucoside with an ion m/z value of 773 was
present in decreasing quantities going from V2 to V5 for
PM09.960 and V2V4 for PM09.812. This latter compound was
identied as a product fragment of the parent molecular ion m/z
of 935 (cyanidin 3-caffeoyl-sophoroside-5-glucoside) indicating
that the parent compound underwent deglycosylation during
digestion. Both these anthocyanins have been detected previously
in cooked purple sweet potatoes (Truong et al., 2010).
Peonidin
3-caffeoyl-p-hydroxy-benzoyl-sophoroside-5-glucoside
(m/z
1069) and its deacylated form peonidin 3-p-hydroxy-benzoyl-so
phoroside-5-glucoside (m/z 907) were found in the stomach V1
containing PM09.960. Peonidin 3-(600 -feruloyl-sophoroside)-5-glu
coside was identied for PM09.960 in the small intestine vessel
V2 based on a molecular ion m/z value of 963 and this anthocyanin
was also found as the fragmentation ion of m/z 801 in the V2V4
vessels for PM09.960, and in V3 and V4 for PM09.812. This anthocyanin has previously been identied in cooked purple sweet potatoes (Truong et al., 2010) and so appears to survive digestion until
the transverse colon. Cyanidin 3-(600 ,600 -dicoumaryl-sophorosid
e)-5-glucoside (m/z 1065) was found only in the small intestine
V2 vessel for PM09.960.
The presence of some of the above anthocyanidin sophorosides
in V2 for both the PM09.812 and PM09.960 accessions could have
biological signicance as anthocyanidin sophorosides and their
derivatives have been demonstrated to exhibit strong
a-glucosidase inhibitory effects that might limit postprandial glucose absorption and so provide anti-diabetic properties (Matsui
et al., 2001). Interestingly, two of the cyanidin species detected
in the colonic vessels, cyanidin 3-caffeoyl-p-hydroxy-benzoylsophoroside-5-glucoside (m/z 1055) in V1 (PM09.960) and 3-caf
feoyl-sophoroside-5-glucoside (m/z 935) in V3 (PM09.812) and
V2 (PM09.960) were noted to be the top two of the three anthocyanins seen in the cooked purple-eshed P40 sweet potato shown
to prevent colorectal cancer in a murine model (Lim et al., 2013; Xu
et al., 2013).
3.2. Antioxidant activity
Antioxidant activity was measured in the 5 vessels of the GI
model before addition of the meal with the sweet potato samples
(T = 0) and in samples withdrawn at 8 h and 24 h (Fig. 3). No major
change in antioxidant capacity was noted in any of the GI vessels
upon addition of the meal that did not contain the sweet potato
samples (T = 0). Upon addition of PM09.812 and PM09.960 sweet
potato meals, an increase in antioxidant capacity was observed at
T = 8 and 24 h in all 5 vessels. It is noteworthy that the antioxidant
power of PM09.812 is generally higher than that of PM09.960. In
addition, as seen in Table 2, the total anthocyanin content of
PM09.812 is substantially higher than that of PM09.960 in V1, an
observation that is mirrored in the difference in FRAP values for
both accessions in V1. For the PFSP accession PM09.960, a spike
in antioxidant activity was observed at 24 h in V2. This increase
also coincides with an increase in total anthocyanin number and
concentrations (Table 2). This observation is likely due to an
increased release of anthocyanins and metabolites or other antioxidant compounds from the food matrix as it enters V2, under the
action of pancreatic enzymes and a drastic pH increase. Dilution
of the parent anthocyanins was likely responsible for the observed
drop in antioxidant capacity seen in V2 versus stomach V1 in
PM09.812. Another major drop in antioxidant capacity was
observed in V3 for both accessions, which was probably the result
of greater diminished concentrations of parent compounds

176

S. Kubow et al. / Food Chemistry 192 (2016) 171177

Sweet Potato PM09.812

FRAP (
g/mL FeS04 equiv.)

14.00
12.00
10.00
8.00

0h (Ctrl)

6.00

8h
24h

4.00
2.00
0.00
V1

V3

V4

V5

Sweet Potato PM09.960

14.00

FRAP (g/mL FeS04 equiv.)

V2

12.00
10.00
8.00

0h (Ctrl)

6.00

8h
24h

4.00
2.00
0.00
V1

V2

V3

V4

V5

Fig. 3. Time course of antioxidant capacity FRAP measures of digesta of gut model
vessels following provision of a meal containing sweet potato PM09.812 and
PM09.960. Data are mean SD. V1 = stomach; V2 = small intestine; V3 = ascending
colon; V4 = transverse colon; V5 = descending colon.

resulting from gut microbial metabolism to generate secondary

metabolites. For the PFSP accession PM09.812, similar drops in
antioxidant activity were observed across the vessels of the gut
model; however, the FRAP values were lower for V1 and V2 at
24 h than at 8 h, a different observation than for accession
PM09.960, suggesting a degradation of antioxidant compounds
within the GI nutrient solution over the length of the digestion
experiment (24 h).
The increase in antioxidant capacity that was observed is supportive of the suggestion that unabsorbed anthocyanins as well
as their microbial metabolites can contribute to antioxidant capacity of human fecal uid. An assessment of the total antioxidant
activity of feces obtained from 14 healthy volunteers showed a
high radical scavenging antioxidant capacity that was almost
20-fold greater than that detected in the plasma (Garsetti,
Pellegrini, Baggio, & Brighenti, 2000). The production of secondary
metabolites from gut microbial metabolism of polyphenols could
exert signicant post-absorptive physiological effects. For example, only microbial metabolites and none of the parent/original
polyphenols in pomegranate juice were detected in the plasma
or urine of healthy volunteers after they consumed pomegranate
juice daily for 5 days (Cerd, Espn, Parra, Martnez, &
Toms-Barbern, 2004). Unabsorbed anthocyanin compounds have
been implicated with enhanced colorectal health and protection
against colorectal cancer via antioxidant mechanisms (Lala et al.,
2006).
4. Concluding remarks
To our knowledge, this is the rst study to report regarding the
effects of digestion on anthocyanins via the human simulated
dynamic in vitro multistage continuous digestion model involving
the ascending colon, transverse colon and descending colon vessels. Although some anthocyanins appear to be stable under simulated upper intestinal tract conditions in the present work, several
anthocyanins were unstable in the intestinal vessel. This

observation is likely due to their chemical decomposition at neutral pH (Woodward, Kroon, Cassidy, & Colin, 2009), which is known
to occur before their subsequent exposure to colonic microbial
metabolism and rst pass intestinal and hepatic metabolism
(Fang, 2014). Anthocyanins are well known to be subject to microbial degradation (Kim et al., 1998). Thus, microbial metabolism can
account for the anthocyanin degradation in the present study as
seen by diminished concentrations of several anthocyanin species
in the colonic vessels, particularly evident in V5. Accession differences in the appearance of anthocyanin species were evident as
PM09.812 showed the appearance of a greater number of anthocyanin species in V3, whereas V2 showed the greatest presence
of anthocyanin species for the PM09.960 accession. This nding
could be indicative of cultivar differences in other food components that might affect anthocyanin bioaccessibility. The two
accessions showed a relatively low quantity of anthocyanin species
noted in the descending colon vessel V5, which could signify
extensive anthocyanin breakdown in the multistage reactor digestion model used in this study which is an advantage compared to
single batch reactors where there is accumulation of microbial
metabolites with prolonged incubation.
Both sweet potato samples demonstrated an increase in FRAP
antioxidant capacity measures over a 24 h time course, which is
supportive of the concept that antioxidant activities of unabsorbed
anthocyanins and their metabolites protect intestinal cells against
reactive oxygen species (ROS) generated within the gut and attenuate ROS-mediated gut inammatory conditions (Glvez et al.,
2001).
In summary, the present study shows that the release of anthocyanins from the food matrix of sweet potatoes under simulated
gastric, intestinal and colonic digestions is an accession dependent
process. There are signicant losses in the variety of anthocyanin
species detected as digestion proceeds from the intestine and
colon, which differs according to the sweet potato accession.
These ndings suggest that cultivar-based variations in other food
components can affect anthocyanin release during digestive processes, which can also impact on their antioxidant capacities in
the small intestine.
Conict of interest
The authors declare that they have no conict of interest in the
research.
Acknowledgments
This research was supported by the CGIAR Research Program on
Agriculture for Nutrition & Health (CRP-A4NH) and the Discovery
Grant Program from the Natural Sciences and Engineering
Council of Canada (S.K.). The authors would like to thank Dr.
Wolfgang Grneberg (CIP) for providing the accessions, Dr.
Gordon Prain (CIP) for enabling the collaboration between CIP
and McGill University, and Mr. Federico Diaz, M.Sc. (CIP) for performing the eld trial in Peru.
References
Alminger, M., Aura, A. M., Bohn, T., Dufour, C., El, S. N., Gomes, A., et al. (2014). In
vitro models for studying secondary plant metabolite digestion and
bioaccessibility. Comprehensive Reviews in Food Science and Food Safety, 13,
413436.
Aura, A.-M., Martin-Lopez, P., OLeary, K. A., Williamson, G., Oksman-Caldentey, K.M., Poutanen, K., et al. (2005). In vitro metabolism of anthocyanins by human
gut microora. European Journal of Nutrition, 44, 133142.
Benzie, I. F. F., & Strain, J. J. (1996). The ferric reducing ability of plasma (FRAP) as a
measure of antioxidant power: The FRAP assay. Analytical Biochemistry, 239,
7076.

S. Kubow et al. / Food Chemistry 192 (2016) 171177

Blanquet-Diot, S., Deat, E., Jarrige, J. F., Denis, S., Beyssac, E., & Alric, M. (2009).
Combining the dynamic TNO-gastrointestinal tract system with a Caco-2 cell
culture model: Application to the assessment of lycopene and alpha-tocopherol
bioavailability from a whole food. Journal of Agricultural and Food Chemistry,
57(23), 1131411320.
Cerd, B., Espn, J. C., Parra, S., Martnez, P., & Toms-Barbern, F. A. (2004). The
potent in vitro antioxidant ellagitannins from pomegranate juice are
metabolised into bioavailable but poor antioxidant hydroxy-6Hdibenzopyran-6-one derivatives by the colonic microora of healthy humans.
European Journal of Nutrition, 43, 205220.
Clifford, M. N. (2000). Anthocyanins Nature, occurrence and dietary burden.
Journal of the Science of Food and Agriculture, 80, 10631072.
Fang, J. (2014). Some anthocyanins could be efciently absorbed across the
gastrointestinal mucosa: Extensive presystemic metabolism reduces apparent
bioavailability. Journal of Agricultural and Food Chemistry. http://dx.doi.org/
10.1021/jf405356b.
Glvez, J., Coelho, G., Crespo, M. E., Cruz, T., Rodrguez-Cabezas, M. E., Concha, A.,
et al. (2001). Intestinal anti-inammatory activity of morin on chronic
experimental colitis in the rat. Alimentary Pharmacology & Therapeutics, 15,
20272039.
Garsetti, M., Pellegrini, N., Baggio, C., & Brighenti, F. (2000). Antioxidant activity in
human faeces. British Journal of Nutrition, 84, 705710.
Kano, M., Takayanagi, T., & Harada, K. (2005). Antioxidative activity of anthocyanins
from purple sweet potato, Ipomoea batatas cultivar Ayamurasaki. Bioscience,
Biotechnology and Biochemistry, 69(5), 979988.
Kim, D. H., Jung, E. A., Sihng, I. S., Han, J. A., Kim, T. H., & Han, M. J. (1998). Intestinal
bacterial metabolism of avonoids and its relation to some biological activities.
Archives of Pharmacal Research, 21, 1723.
Kim, H. W., Kim, J. B., Cho, S. M., Chung, M. N., Lee, Y. M., Chu, S. M., et al. (2012).
Anthocyanin changes in the Korean purple-eshed potato Shinzami, as affected
by steaming and baking. Food Chemistry, 130, 966972.
Kumarathasan, P., Vincent, R., Das, D., Mohottalage, S., Blais, E., Blank, K., et al.
(2014). Applicability of a high-throughput shotgun plasma protein screening
approach in understanding maternal biological pathways relevant to infant
birth weight outcome. Journal of Proteomics, 100, 136146.
Lala, G., Malik, M., Zhao, C., He, J., Kwon, Y., Giusti, M. M., et al. (2006). Anthocyaninrich extracts inhibit multiple biomarkers of colon cancer in rats. Nutrition and
Cancer, 54, 8493.
Lee, M. J., Park, J. S., Choi, D. S., & Jung, M. Y. (2013). Characterization and
quantitation of anthocyanins in purple-eshed sweet potatoes cultivated in
Korea by HPLC-DAD and HPLCESI-QTOF-MS/MS. Journal of Agricultural and
Food Chemistry, 61, 31483158.
Lim, S., Xu, J., Kim, J., Chen, T.-Y., Su, X., Standard, J., et al. (2013). Role of
anthocyanin-enriched purple-eshed sweet potato p40 in colorectal cancer
prevention. Molecular Nutrition and Food Research, 57, 19081917.
Martoni, C., Bhathena, J., Jones, M. L., Urbanska, A. M., Chen, H., & Prakash, S. (2007).
Investigation of microencapsulated BSH active Lactobacillus in the simulated
human GI tract. Journal of Biomedicine and Biotechnology, 7, 19.
Matsui, T., Ueda, T., Oki, T., Sugita, K., Terahara, N., & Matsumoto, K. (2001). Alphaglucosidase inhibitory action of natural acylated anthocyanins. 2. Alpha-

177

glucosidase inhibition by isolated acylated anthocyanins. Journal of

Agricultural and Food Chemistry, 49, 19521956.
McDougall, G. J., Dobson, P., Smith, P., Blake, A., & Stewart, D. (2005). Assessing
potential bioavailability of raspberry anthocyanins using an in vitro digestion
system. Journal of Agricultural and Food Chemistry, 53, 58965904.
Molly, K., Vande Woestyne, M., De Smet, I., & Verstraete, W. (1994). Validation of
the simulator of the human intestinal microbial ecosystem (SHIME) reactor
using microorganism-associated activities. Microbial Ecology in Health and
Disease, 7, 191200.
Montilla, E. C., Hillebrand, S., Butschbach, D., Baldermann, S., Watanabe, N., &
Winterhalter, P. (2010). Preparative isolation of anthocyanins from Japanese
purple sweet potato (Ipomoea batatas L.) varieties by high-speed countercurrent
chromatography. Journal of Agricultural and Food Chemistry, 58, 98999904.
Montilla, E. C., Hillebrand, S., & Winterhalter, P. (2011). Anthocyanins in purple
sweet potato (Ipomoea batatas L.) varieties. Fruit, Vegetable and Cereal Science
and Biotechnology, 5, 1924.
Mori, M., Hayashi, K., Ohara-Takada, A., Watanuki, H., Katahira, R., Ono, H., et al.
(2010). Anthocyanins from skins and eshes of potato varieties. Food Science
and Technology Research, 16, 115122.
Mullen, W., Edwards, C. A., Serani, M., & Crozier, A. (2008). Bioavailability of
pelargonidin-3-O-glucoside and its metabolites in humans following the
ingestion of strawberries with and without cream. Journal of Agricultural and
Food Chemistry, 56, 713719.
Pods?dek, A., Redzynia, M., Klewicka, E., & Koziokiewicz, M. (2014). Matrix effects
on the stability and antioxidant activity of red cabbage anthocyanins under
simulated gastrointestinal digestion. BioMed Research International. http://
dx.doi.org/10.1155/2014/365738.
Tarko, T., Duda-Chodak, A., & Zajac, N. (2013). Digestion and absorption of phenolic
compounds assessed by in vitro simulation methods. A review. Roczniki
stwowego Zakadu Higieny, 64(2), 7984.
Pan
Tian, Q., Konczak, I., & Schwartz, S. (2005). Probing anthocyanin proles in purple
sweet potato cell line (Ipomea batatas L. cv. Ayamurasaki) by high performance
liquid chromatography and electrospray ionization tandem mass spectrometry.
Journal of Agricultural and Food Chemistry, 53, 65036509.
Truong, V. D., Deighton, N., Thompson, R. L., McFeeters, R. F., Dean, L. O., Pecota, K.
V., et al. (2010). Characterization of anthocyanins and anthocyanidins in purpleeshed sweetpotatoes by HPLC-DAD/ESI-MS/MS. Journal of Agricultural and Food
Chemistry, 58, 404410.
Woodward, G., Kroon, P., Cassidy, A., & Colin, K. (2009). Anthocyanin stability and
recovery: Implications for the analysis of clinical and experimental samples.
Journal of Agricultural and Food Chemistry, 57, 52715278.
Wu, L.-H., Xu, Z.-L., Dong, D., He, S.-A., & Yu, H. (2011). Protective effect of
anthocyanins extracts from blueberry on TNBS-induced IBD model of mice.
Evidence Based Complementary Alternative Medicine. http://dx.doi.org/10.1093/
ecam/neq040.
Xu, J., Su, X., Lim, S., Grifn, J., Carey, E., Katz, B., et al. (2013). Identication of
anthocyanins in purple-eshed sweetpotato and stability during various
cooking conditions. The FASEB Journal, 27(636), 23.