Carotenoids and carotenoid esters in potatoes (Solanum tuberosum L.): New insights into an ancient vegetable

Article abstract of DOI:10.1021/jf0257953

The carotenoid pattern of four yellow- and four white-fleshed potato cultivars (Solanum tuberosum L.), common on the German market, was investigated using HPLC and LC(APCI)-MS for identification and quantification of carotenoids. In each case, the carotenoid pattern was dominated by violaxanthin, antheraxanthin, lutein, and zeaxanthin, which were present in different ratios, whereas neoxanthin, β-cryptoxanthin, and β,β-carotene generally are only minor constituents. In contrast to literature data, antheraxanthin was found to be the only carotenoid epoxide present in native extracts. The total concentration of the four main carotenoids reached 175,ug/100 g, whereas the sum of carotenoid esters accounted for 41-131 μg/100 g. Therefore, carotenoid esters are regarded as quantitatively significant compounds in potatoes. For LC(APCI)-MS analyses of carotenoid esters, a two-stage cleanup procedure was developed, involving column chromatography on silica gel and enzymatic cleavage of residual triacylglycerides by lipases. This facilitated the direct identification of several potato carotenoid esters without previous isolation of the compounds. Although the unequivocal identification of all parent carotenoids was not possible, the cleanup procedure proved to be highly efficient for LC(APCI)-MS analyses of very low amounts of carotenoid esters.

Full text of DOI:10.1021/jf0257953

J. Agric. Food Chem. 2002, 50, 7175−7181 7175
Carotenoids and Carotenoid Esters in Potatoes (Solanum
tuberosum L.): New Insights into an Ancient Vegetable
DIETMAR E. BREITHAUPT* AND AMENEH BAMEDI
Institut fu¨r Lebensmittelchemie, Universita¨t Hohenheim, Garbenstrasse 28,
D-70593 Stuttgart, Germany
The carotenoid pattern of four yellow- and four white-fleshed potato cultivars (Solanum tuberosum
L.), common on the German market, was investigated using HPLC and LC(APCI)-MS for identification
and quantification of carotenoids. In each case, the carotenoid pattern was dominated by violaxanthin,
antheraxanthin, lutein, and zeaxanthin, which were present in different ratios, whereas neoxanthin,
â-cryptoxanthin, and â,â-carotene generally are only minor constituents. In contrast to literature data,
antheraxanthin was found to be the only carotenoid epoxide present in native extracts. The total
concentration of the four main carotenoids reached 175 µg/100 g, whereas the sum of carotenoid
esters accounted for 41-131 µg/100 g. Therefore, carotenoid esters are regarded as quantitatively
significant compounds in potatoes. For LC(APCI)-MS analyses of carotenoid esters, a two-stage
cleanup procedure was developed, involving column chromatography on silica gel and enzymatic
cleavage of residual triacylglycerides by lipases. This facilitated the direct identification of several
potato carotenoid esters without previous isolation of the compounds. Although the unequivocal
identification of all parent carotenoids was not possible, the cleanup procedure proved to be highly
efficient for LC(APCI)-MS analyses of very low amounts of carotenoid esters.
KEYWORDS: Potato; Solanum tuberosum; carotenoid ester; carotenoid epoxide
INTRODUCTION
eight other vegetables and five fruitssto the foods accounting
for more than 96% of the intake of the major carotenoids in
human serum, both on a yearly and on a seasonal basis (7).
Potatoes (Solanum tuberosum L.) are grown in many countries
with temperate climate all over the world. Today, potatoes are
among the most commonly consumed vegetables and serve as
a major source of vitamin C. Carotenoids, especially â,â-
carotene, represent minor constituents. Hence, potatoes are not
an important source of provitamin A (1). However, lutein, one
of the main carotenoids, has attracted interest since high serum
levels are correlated with a reduced risk for humans to be
afflicted with age-related macular degeneration (AMD) (2, 3).
Since typical green and leafy vegetables (e.g. spinach, broccoli),
the major dietary sources of lutein, are not frequently consumed,
potatoes serve as additional source of lutein in the diet (4).
Concerning the health benefits of other typical potato carotenoids
(e.g. violaxanthin, neoxanthin, antheraxanthin), only little is
known. With respect to the bioavailability of epoxy-carotenoids,
a recently published study ascertained that epoxy-xanthophylls
(e.g. violaxanthin) are not absorbed by humans (5). However,
due to their widespread acceptability, potatoes are generally
expected to be a valuable source of carotenoids in the diet,
especially if intensely colored clones are selected (4) or modern
breeding techniques are applied (6). The contribution of potatoes
as a plant source of carotenoids was demonstrated by a study
conducted in 1996 in Spain: potato belongedstogether with
With respect to potato carotenoids, qualitative as well as
quantitative data in the literature are extremely confusing.
Kasim, one of the first researchers who reported on these
micronutrients in potatoes, used paper chromatography and
visible spectroscopy for identification of several carotenoids in
saponified potato extracts (8, 9). Besides lutein and violaxanthin,
he detected lutein-5,6-epoxide and â,â-carotene diepoxide,
whereas zeaxanthin was not found. Pendlington et al. investi-
gated the carotenoid pattern using alumina columns to separate
carotenoids and tentatively identified eight carotenoids (e.g. â,â-
carotene, lutein) as most abundant in all cultivars (10). Mu¨ller
reported a carotenoid screening of various vegetables, applying
chemical saponification before HPLC analyses (11). According
to these results, violaxanthin is the main potato carotenoid,
followed by lutein, antheraxanthin, and others. Iwanzik et al.
(1) and Stute et al. (12) compared the carotenoid content of 13
potato cultivars from the same habitat. The main carotenoid was
again violaxanthin, followed by lutein, lutein-5,6-epoxide,
neoxanthin, and neoxanthin a; zeaxanthin was not identified.
The first to mention the occurrence of carotenoid esters in
amyloplast envelope membranes, the site of localization of
potato carotenoids, were Fishwick and Wright (13). Although
they did not attempt to identify the fatty acid component of the
carotenoid esters, they reported 17-22% of the carotenoids to
* To whom correspondence should be addressed [telephone (49)
7114594094; fax (49) 7114594096; E-mail breithau@uni-hohenheim.de].
10.1021/jf0257953 CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/25/2002

7176 J. Agric. Food Chem., Vol. 50, No. 24, 2002
Breithaupt and Bamedi
Turrax T 25 (Janke & Kunkel). Aliquots of 30-50 g were immediately
extracted with light petroleum/methanol/ethyl acetate (1:1:1 v/v/v; 4
× 100 mL) until the extracts were colorless (total volume: 1.5 L).
The upper layer was separated; the lower layer was shaken with water
(100 mL) for phase separation and the upper layer combined with the
first extract (total volume: 800 mL). The organic extract was dried
with anhydrous sodium sulfate (30 g), filtered through a folded filter,
and evaporated to dryness in a vacuum at 30 °C, and the residue was
dissolved in light petroleum (5 mL). For determination of the native
carotenoid pattern, an aliquot (1 mL) was evaporated in a gentle stream
of nitrogen, and the dry residue was redissolved in MTBE/methanol/
BHT (1:1:0.01 v/v/w; 1 mL) and subjected to HPLC analysis. All
procedures were done under dim light.
For saponification, another aliquot of the light petroleum solution
(3 mL) was transferred to a flat-bottomed flask, dried in a stream of
nitrogen, dissolved in diethyl ether (100 mL), and saponified at room
temperature overnight with methanolic potassium hydroxide (30% w/v;
1 mL). For complete removal of alkali, the solution was washed twice
with distilled water (100 mL each). The organic layer was dried over
anhydrous sodium sulfate (20 g), filtered through a folded filter, and
evaporated to dryness, and the residue was redissolved in MTBE/
methanol/BHT (1:1:0.01 v/v/w; 2 mL) and subjected to HPLC analysis.
For quantitative determination, peak areas of carotenoids in the sample
chromatograms were correlated with the concentrations on the basis
of calibration curves, established as detailed earlier (18). UV spectra
were usually recorded in the HPLC eluent by DAD. The color of the
flesh (white/yellow) was evaluated by visual examination of freshly
sliced tubers.
Synthesis of Lutein, Zeaxanthin, and â,â-Carotene Epoxides.
Carotenoid epoxides were synthesized on the basis of a method
described by Barua (19, 20) and Barua and Olson (5), comprising the
chromatographic separation of the products. Lutein and zeaxanthin (1
mg each), obtained from saponified marigold oleoresin (Tagetes erecta
L.) or saponified extracts of orange pepper (Capsicum annuum L.),
and â,â-carotene, respectively, were dissolved in diethyl ether (10 mL).
After a solution of 3-chloroperoxybenzoic acid (5 mg/10 mL) was
added, each mixture was stirred at room temperature in the dark for
16 h. After the mixture was washed once with sodium carbonate
solution (0.1 M, 50 mL) and twice with water (50 mL each), the ether
layer was separated, dried over anhydrous sodium sulfate (2 g), filtered,
and evaporated to dryness in a rotary evaporator at 30 °C. Negative
starch iodine reaction in the washing water indicated complete removal
of peroxides. The residue was dissolved in MTBE/methanol/BHT (1:
1:0.01 v/v/w; 5 mL), membrane filtered, and subjected to HPLC and
LC(APCI)-MS analyses.
be present in the esterified form. As main parent carotenoids,
antheraxanthin, neoxanthin, lutein/zeaxanthin, and â-cryptox-
anthin were characterized. Basic work in the field of potato
carotenoid ester analysis has also been carried out by Tevini
and co-workers (14, 15). They separated mono- and diester
fractions and tentatively identified carotenoid esters (e.g. lutein-
5,6-epoxide mono- and diesters) by comparing their visible
spectra. Unfortunately, the individual fatty acid moiety was not
investigated. However, Tevini’s group analyzed the fatty acid
pattern of the total carotenoid ester fraction and found palmitic,
linoleic, linolenic, and stearic acid to be the main components.
Recently, Lu et al. presented the carotenoid content of 11 diploid
potato clones and two tetraploid cultivars (4). Although they
used no saponification step, the existence of carotenoid esters
was not mentioned, probably leading to an underestimation of
the respective carotenoids. The same is true for a study dealing
with orange-fleshed potatoes, establishing that the color origi-
nates from large amounts of zeaxanthin (16). Since saponifica-
tion was used prior to analysis, no information about the
occurrence of zeaxanthin esters was obtained.
These inconsistent data may be traced back, on one hand, to
unsuitable isolation and identification techniques and, on the
other, to degradation or rearrangement reactions during the
workup procedure. The lack of data concerning zeaxanthin may
reflect the fact that, so far, it has been difficult to separate lutein
from zeaxanthin. However, until now, little data were available
about the carotenoid content of potato, especially with respect
to carotenoid esters, which have not been studied using modern
analytical techniques in the past 20 years. Therefore, we
investigated the carotenoid content of four yellow- (Granola,
Solara, Marabel, and Nicola) and four white-fleshed (Siglinde,
Sante, Bintje, and Cilena) cultivars, common on the German
market. For identification of free and esterified carotenoids, LC-
(APCI)-MS analyses in the positive mode were applied. The
carotenoid pattern of sweet potato (Ipomoea batatas (L.) Lam.),
common in tropical areas, is dominated by â,â-carotene and
â,â-carotene mono- and diepoxides and was not investigated
in this study (17).
MATERIALS AND METHODS
Chemicals. Light petroleum (boiling fraction 40-60 °C), diethyl
ether, methanol, butylated hydroxytoluene (BHT), ethyl acetate, and
calcium carbonate were purchased from Merck (Darmstadt, Germany),
and 3-chloroperoxybenzoic acid (∼70%), â,â-carotene, sodium sulfate
(anhydrous), sodium carbonate, potassium hydroxide, and methyl tert-
butyl ether (MTBE) were purchased from Sigma-Aldrich GmbH
(Taufkirchen, Germany). All solvents were distilled before use. High-
purity water was prepared with a Milli-Q 185 Plus water purification
system (Millipore, Eschborn, Germany). â-Cryptoxanthin was gener-
ously provided by Hoffmann-La Roche (Basel, Switzerland). Neox-
anthin (c ) 1.137 mg/L) and violaxanthin (c ) 0.637 mg/L; certified
standard solutions in ethanol) were purchased from DHI-
Water&Environment (Hørsholm, Denmark). Novozym 868 L (6 kilo
lipase units (KLU)/g) and Lipolase (100 KLU/g) were kindly supplied
by Novo Nordisk Biotechnology GmbH (Mainz, Germany). Fresh
orange peppers (for extraction of zeaxanthin) were obtained from the
local market, and marigold oleoresin (for extraction of lutein) was
supplied by Euram Food GmbH (Stuttgart, Germany). An extract of a
genetically transformed Escherichia coli strain (containing the two
plasmids pACCAR25delcrtX and pBBR-zep), producing antheraxanthin
and zeaxanthin in high yields, was kindly supplied by G. Sandmann,
Botanical Institute, J.W. Goethe University (Frankfurt, Germany).
Cleanup of Native Extracts for LC(APCI)-MS Analysis. Four
extracts obtained as described above were pooled and evaporated to
dryness in a rotary evaporator, and the residue was dissolved in light
petroleum (5 mL) and subjected to column chromatography on silica
gel as described earlier (21). Carotenoid esters were eluted using light
petroleum containing different amounts of acetone (2, 5, and 10%
(v/v)). Thus, polar carotenoids remained on the column and were
discarded. For further purification, an enzymatic cleanup step was
applied (18). In brief, samples cleaned by column chromatography were
transferred into a sealable glass tube, the solvent was evaporated in a
gentle stream of nitrogen, and 0.1 M phosphate buffer pH 7.4 (10 mL)
was added. Bile salts and sodium/calcium ions were used as activators.
For triacylglyceride hydrolysis, an aliquot (200 µL) of a freshly prepared
lipase suspension containing Novozym 868 L (20 µL) and Lipolase
(20 µL) in 5 mM calcium chloride solution (1 mL) was added, and the
mixture was incubated at 37 °C for 2 h. Afterwards, the carotenoids
were extracted using methanol/ethyl acetate/light petroleum (1:1:1
v/v/v; 3 × 20 mL), processed as described above, and subjected to
LC(APCI)-MS analysis.
Recovery Studies. For recovery experiments, a solution of â-cryp-
toxanthin in MTBE/methanol (1:1 v/v) was prepared (100 µg/mL) and
stored at -18 °C. Aliquots (1 mL) were added to minced potatoes
(variety Granola; 150 g each), low in native â-cryptoxanthin content,
and immediately subjected to the workup procedure. Thus, the resulting
â-cryptoxanthin content of fortified samples (67 µg/100 g) lies in the
typical range of the native concentration of a carotenoid in potato
Preparation of Samples. Potato tubers, obtained from local markets
in the autumn of 2001 (grade of goods: I), were peeled and cut into
small cubes. An aliquot (150 g; three to five tubers) was mixed with
solid calcium carbonate (5 g) and homogenized for 2 min by an Ultra

Carotenoids in Potatoes
J. Agric. Food Chem., Vol. 50, No. 24, 2002 7177
Figure 1. Typical chromatograms (DAD, 450 nm; retention time window 15−45 min) of a native (A) and a saponified (B) white-fleshed potato extract
(cultivar Bintje). Peak identification: 1, neoxanthin; 2, violaxanthin; 3, unidentified; 4, antheraxanthin; 5, lutein; 6, zeaxanthin; 7, â-cryptoxanthin; 8,
â,â-carotene. Due to the workup procedure, the saponified sample (B) is concentrated by a factor of 1.5.
samples. Recoveries were calculated according to AOAC methods (22)
as follows: % Recovery ) [(cA - cU)/cB] × 100, where cA is the
concentration of â-cryptoxanthin measured in the fortified sample, cU
is the concentration of â-cryptoxanthin measured in the unfortified
sample, and cB is the concentration of â-cryptoxanthin added in fortified
samples. The following recoveries were obtained (n ) 4): without
saponification, 83.4 ( 2.7%; after saponification, 81.5 ( 4.8%. These
values are regarded as acceptable since they lie between a range of
80-110%, which is accepted for concentrations of approximately 1
mg/kg by regulatory agencies in the United States (23).
HPLC and LC(APCI)-MS Analyses. Analyses were performed
with a modular HPLC system HP1050 (Hewlett-Packard, Waldbronn,
Germany) and a column thermoregulator (Mistral, Spark, The Neth-
erlands). A 250 mm × 4.6 mm i.d., 5 µm YMC C30 reverse-phase
column from YMC Europe (Schermbeck, Germany), equipped with a
10 mm × 4.6 mm i.d., 5 µm Nucleosil C18 precolumn (Bischoff,
Leonberg, Germany), was used and kept at 35 °C. Two mobile-phase
systems were used. For HPLC analyses, system I was used: mobile
phases methanol/water/triethylamine (90:10:0.1 v/v/v) (A) and methanol/
MTBE/water/triethylamine (6:90:4:0.1 v/v/v/v) (B); gradient (min/%A)
0/99, 8/99, 45/0, 50/99, and 55/99. For analysis of carotenoid esters
by LC(APCI)-MS, system II was applied: mobile phases methanol/
MTBE/water/triethylamine [81:15:4:0.1 v/v/v/v (A) and 6:90:4:0.1 v/v/
v/v (B)]; gradient (min/%A) 0/99, 39/44, 45/0, 50/99, and 55/99. The
flow rate was 1 mL/min, the injection volume 20 µL, and the detection
wavelength (DAD) 450 nm in both systems. LC(APCI)-MS was run
on an HP1100 HPLC system, coupled to a Micromass (Manchester,
UK) VG platform II quadrupole mass spectrometer equipped with an
APCI+ interface; the MS parameters have been previously reported
(24).
Figure 2. Structures of the most important carotenoids found in potatoes.
The numbering corresponds to the carotenoids indicated in Figure 1.
RESULTS AND DISCUSSION
Qualitative and Quantitative Results. Since violaxanthin,
a carotenoid which is rather susceptible to acidic conditions,
may occur in potatoes, we mixed the potato paste immediately
after homogenization with solid calcium carbonate to raise the
pH of the mixture. As preliminary tests showed, this procedure,
suggested previously by Lu et al. (4), enabled us to determine
native violaxanthin without the risk of destruction in potato
extracts. We investigated the carotenoid content of four yellow-
(Granola, Solara, Marabel, and Nicola) and four white-fleshed
(Siglinde, Sante, Bintje, and Cilena) cultivars, common in
Germany. The carotenoid pattern of native extracts, shown in
Figure 1A, was dominated by violaxanthin (2), antheraxanthin
(4), lutein (5), and zeaxanthin (6) in different ratios, followed
by â-cryptoxanthin (7) and â,â-carotene (8), being only byprod-
ucts; neoxanthin (1), a carotenoid eluting just in front of
violaxanthin, represented in most cases another minor constituent
(for structures of the carotenoids, see Figure 2). The identifica-
tion of 1, 2, 5, 6, 7, and 8 was confirmed by comparison of
their retention times and UV spectra with those of commercial
authentic reference material; UV data were consistent with those
reported (25), as well as the fragmentation pattern in LC-
(APCI)-MS analyses. In principle, these results are in ac-
cordance with those of Mu¨ller (11), who found violaxanthin to
be the main carotenoid (180 µg/100 g), followed by mutato-/

7178 J. Agric. Food Chem., Vol. 50, No. 24, 2002
Breithaupt and Bamedi
Table 1. Concentration of Carotenoids [µg/100 g Fresh Weight] of Different Potato Cultivars (Edible Part), Concentration of Carotenoid Esters,
Calculated as Lutein Dimyristate, and Color of the Flesh
carotenoids [µg/100 g]a
cultivar
color
violaxanthin
antheraxanthinb
lutein
zeaxanthin
â-cryptoxanthin
all-trans-â-carotin
diestersc
Granola
yellow
66.3 ± 13.6
70.6 ± 11.2
15.0 ± 3.3
23.8 ± 2.5
7.8 ± 1.6
40.6 ± 7.0
9.1 ± 1.8
27.4 ± 2.3
12.8 ± 2.0
24.5 ± 2.1
7.7 ± 1.7
19.7 ± 2.5
3.3 ± 0.7
6.0 ± 0.2
4.9 ± 1.2
20.3 ± 2.4
28.8 ± 5.4
33.5 ± 2.1
27.9 ± 2.8
30.6 ± 2.3
48.0 ± 7.0
66.1 ± 4.1
21.6 ± 3.5
41.8 ± 4.6
11.8 ± 1.3
17.8 ± 0.9
7.7 ± 2.3
26.7 ± 4.5
27.3 ± 4.4
27.0 ± 2.6
32.0 ± 3.8
40.5 ± 5.0
48.9 ± 4.9
16.8 ± 1.3
24.4 ± 1.8
20.7 ± 2.0
23.8 ± 2.6
20.8 ± 1.7
22.1 ± 0.6
21.0 ± 2.9
24.2 ± 2.3
20.6 ± 4.8
29.5 ± 1.2
8.5 ± 1.3
8.6 ± 2.1
4.9 ± 1.0
4.8 ± 1.5
3.9 ± 0.6
5.8 ± 0.4
4.8 ± 0.7
2.2 ± 0.6
2.5 ± 0.5
1.1 ± 0.1
2.4 ± 0.7
1.8 ± 0.4
2.4 ± 0.5
2.5 ± 0.8
0.8 ± 0.2
0.3 ± 0.1
2.3 ± 0.3
0.5 ± 0.2
1.8 ± 0.4
2.2 ± 0.8
3.4 ± 0.6
3.6 ± 0.8
2.2 ± 0.5
2.3 ± 0.5
2.0 ± 0.6
1.6 ± 0.6
3.4 ± 0.3
2.6 ± 0.9
1.6 ± 0.6
1.4 ± 0.6
1.0 ± 0.2
1.0 ± 0.1
2.4 ± 0.2
2.2 ± 0.2
56.0 ± 5.1
Solara
Marabel
Nicola
Siglinded
Sante
yellow
yellow
yellow
white
white
white
white
20.7 ± 4.0
23.3 ± 1.7
78.3 ± 8.9
107.4 ± 5.5
10.0 ± 1.6
33.6 ± 2.8
6.1 ± 0.8
10.8 ± 0.9
5.6 ± 0.6
6.4 ± 1.1
96.2 ± 4.6
131.2 ± 2.8
116.8 ± 7.2
100.0 ± 7.1
62.6 ± 9.8
40.7 ± 4.8
94.3 ± 8.2
12.2 ± 2.0
20.9 ± 3.8
26.2 ± 3.1
9.8 ± 2.0
Bintje
16.9 ± 5.7
19.6 ± 1.7
2.7 ± 0.8
Cilena
23.7 ± 2.1
16.7 ± 3.6
a The given values represent means ± standard deviations of five independent determinations (based on the workup of different tubers). The first line lists the concentration
of free carotenoids. The second line gives the carotenoid content after saponification. b Antheraxanthin was calculated as zeaxanthin. c Calculated as lutein dimyristate.
d Exact cultivar: Siglinde Hochmoor.
Figure 3. HPLC chromatogram (DAD, 450 nm; retention time window 10−50 min) of a native potato extract (cultivar Nicola) after application of a
two-stage cleanup procedure (for conditions, see Materials and Methods; for precise peak assignment, see Table 2).
antheraxanthin, lutein, zeaxanthin, and neoxanthin; â,â-carotene
and â-cryptoxanthin are minor components. Additionally, caro-
tenoid esters were found in varying amounts in each cultivar.
Analytical proof is provided by the disappearance of the
respective peaks after applying alkaline saponification. As
expected, the carotenoid pattern of the resulting saponified
samples was more complex (Figure 1B) than that of native
samples: on one hand, amounts of xanthophylls already present
in native samples increased, while on the other hand, several
new minor carotenoids appeared in the more polar region. In
all cases, a more or less intense compound (3; m/z 601 [M +
H]+; λmax 414, 436, 466 nm) appeared in front of antheraxanthin
(4). Moreover, the mass spectra of the peak at the retention time
of zeaxanthin (6) showed in some cases (e.g. Nicola, Marabel)
an additional but less intense ion signal (m/z 585). This signal
does definitely not belong to the spectrum of the pure compound
(6), pointing to a coeluting carotenoid. Attempts to separate this
substance by variation of the gradient failed. However, lutein-
5,6-epoxide (see below) can be ruled out as responsible for the
two phenomena mentioned above. Unfortunately, currently an
unequivocal identification of both naturally esterified carotenoids
cannot be provided.
On the basis of the carotenoid pattern observed, it was
difficult to differentiate between white- and yellow-fleshed
cultivars. In our investigations, neither white- nor yellow-fleshed
potatoes showed typical carotenoids as being responsible for
their inherent color. For example, the main carotenoid in yellow-
fleshed cultivars can be antheraxanthin (Solara, Nicola) or
zeaxanthin (Marabel), whereas in Granola violaxanthin domi-
nated the spectrum. The total concentration of carotenoids seems
to be a good tool to differentiate between the two groups: the
sum of the four main carotenoids determined in native samples
of yellow-fleshed cultivars was in the range of 58-175 µg/100
g, whereas that of white-fleshed cultivars was between 38 and
62 µg/100 g, respectively. Iwanzik et al. (1) reported concentra-
tions in the range of 171-343 µg/100 g (intensively yellow)
and 27-74 µg/100 g (white), verifying the overall tendency
observed in our study. Carotenoid esters, calculated in native
samples as lutein dimyristate, accounted for between 41 µg/

Carotenoids in Potatoes
J. Agric. Food Chem., Vol. 50, No. 24, 2002 7179
Table 2. LC(APCI)−MS Data and Identification of Carotenoid Diesters Originating from the Cultivar Nicola
m/z (relative intensity)a
9
10
11
12
13
14
15
16
[M + H]+
1021
(15)
1003
(67)
775
(100)
547
(30)
793
(12)
565
(5)
418/
438/
466
C14:0
C14:0
violaxanthin
1049
(28)
1031
(91)
803/775
(100/95)
547
1021
(100)
1003
(46)
775
(10)
547
(5)
793
(10)
565
(10)
402/
422/
446
C14:0
C14:0
b
1049
(100)
1031
(54)
803
(7)
547
(5)
821/793
(6/5)
565
1005
(100)
987
(17)
759
(26)
531
(13)
777
(28)
549
(10)
428/
454
989
(100)
[M + H − H O]+
2
[M + H − H O − FA ]+
2
1
[M + H − H O − FA − FA ]+
2
1
2
(32)
[M + H − FA ]+
821/793
(16/22)
565
(15)
418/
438/
466
C14:0
C16:0
violaxanthin
761
(100)
533
761
(47)
533
(5)
420/
452/
480
C14:0
C14:0
zeaxanthin
789/761
(100/67)
533
(57)
417/
444/
471
C14:0
C16:0
lutein
1
[M + H − FA − FA ]+
1
2
(12)
(38)
λmax (DAD)
402/
422/
446
C14:0
C16:0
b
417/
444/
471
C14:0
C14:0
lutein
type of diester
C14:0
C14:0
b
parent carotenoid
a The numbers of the compounds 9−16 correspond to the peak numbering in Figure 3. b The parent carotenoid was not unequivocally identified.
100 g (Bintje) and 131 µg/100 g (Marabel) and, therefore, are
quantitatively significant compounds in the potato carotenoid
spectrum.
monoepoxides are characterized by three maxima (λmax 416,
440, 470 nm), which are consistent with previous data (25).
The twin peaks are probably due to diastereoisomers formed
during epoxidation; the diastereoisomeric profile was not studied
in detail. In the case of zeaxanthin monoepoxides, the quasi-
molecular ion (m/z 585 [M + H]+) formed the base peak,
whereas the fragment ion produced by loss of water was of low
intensity (13%); this is in line with the behavior of zeaxanthin
(m/z 569 [M + H]+ (100%), 551 (7%); λmax 420, 452, 480 nm).
As expected, the spectroscopic characteristics (λmax 419, 446,
474 nm) show a hypsochromic shift compared to zeaxanthin.
Since 4 (Figure 1) showed the same characteristics as zeaxan-
thin monoepoxide, also designated antheraxanthin (3,3′-dihy-
droxy-5,6-epoxy-â,â-carotene), it can be concluded that this
peak corresponds to this compound. For verification, we used
an extract of a genetically modified E. coli strain, producing
antheraxanthin in addition to zeaxanthin. LC(APCI)-MS
analysis of both compounds clearly proved 4 to be antherax-
anthin. For further evidence, one potato extract was evaporated,
and the residue was dissolved in ethanol and treated with a few
drops of diluted hydrochloric acid to convert the presumed 5,6-
epoxide (antheraxanthin) to its 5,8-isomer (mutatoxanthin) (27).
An LC(APCI)-MS analysis proved that the new main peak
which appeared in the chromatogram represents the expected
carotenoid (m/z 585 [M + H]+ (100%), 567 (11%); λmax 405
(sh), 426, 453 nm). The occurrence of antheraxanthin is in
contrast to reports of Iwanzik et al. (1) and Kasim (8, 9), who
found lutein-5,6-epoxide in potato extracts.
Epoxides of â,â-carotene are easily accessible via epoxidation
with 3-chloroperoxybenzoic acid, yielding both mono- and
diepoxides in high amounts. Neither â,â-carotene monoepoxide
(m/z 553 [M + H]+ (100%), 535 (24%); λmax 420, 445, 472
nm) nor â,â-carotene diepoxide (m/z 569 [M + H]+ (100%),
551 (46%); λmax 415, 439, 468 nm) was present in the potato
extracts examined in this study.
LC(APCI)-MS Analyses of Carotenoid Esters. As an
example, four native extracts of the cultivar Nicola, correspond-
ing to 600 g of fresh material, were pooled, purified using
column chromatography on silica gel as described earlier (21),
and analyzed by LC(APCI)-MS. Unfortunately, the presence
of residual triacylglycerides resulted in a high background noise,
which forestalled monitoring of target compound spectra. To
avoid this interference, an additional enzymatic cleanup pro-
The quantitative results of all cultivars investigated are
presented in Table 1. The values given represent means of five
independent determinations. Thus, the standard deviations reflect
the native variation of the carotenoid content of the samples,
obtained from three to five tubers each. The wide variation
observed in some cases (especially if considering very low
concentration carotenoids, such as â-cryptoxanthin and â,â-
carotene) may be the result of this small number of tubers used
for a single determination, the stage of ripeness, and the weather
conditions during growth. The results of recovery studies (native
samples, 83.4 ( 2.7%; after saponification, 81.5 ( 4.8%)
demonstrate that there is only minor loss of major carotenoids
during the extraction process or the saponification procedure,
pointing out that the standard deviations are due to differences
in the carotenoid content of the respective samples. However,
for cases of extremely low concentrated carotenoids (â-
cryptoxanthin and â-carotene; see Table 1), the lower concen-
trations obtained in some cases after saponification may actually
be a result of losses or degradation reactions during this workup
step.
Epoxides in Potato. Special attention was paid to the question
of whether lutein, zeaxanthin, or â,â-carotene epoxides are
present in native extracts. The epoxidation/de-epoxidation
reaction sequence of violaxanthin to zeaxanthin via antherax-
anthin is known as the violaxanthin cycle. Thus, antheraxanthin
represents an intermediate in both directions and can be expected
in plants showing the respective enzymatic activity, e.g. potatoes
and tobacco (26). To confirm this theory, we independently
synthesized the epoxides of lutein, zeaxanthin, and â,â-carotene
and investigated the resulting solutions using LC(APCI)-MS.
The respective chromatograms of lutein and zeaxanthin epoxides
showed four peaks each. MS analyses proved two peaks to
correspond to diepoxy xanthophylls, the others representing
monoepoxides. In the case of lutein monoepoxides, the quasi-
molecular ions (m/z 585 [M + H]+) were of minor intensity
(36%), whereas the fragment ions generated by loss of water
(m/z 567) formed the base peak. This water elimination is typical
for lutein (m/z 569 [M + H]+ (8%), 551 (100%); λmax 417,
444, 471 nm), since the allylic cation generated at the C-3′
position is stabilized by resonance (21). The UV spectra of lutein

7180 J. Agric. Food Chem., Vol. 50, No. 24, 2002
Breithaupt and Bamedi
analyzed. Various other fatty acids may acylate minor caro-
tenoids. Thus, the complete fatty acid pattern of carotenoid esters
is not known. Analyses of the triacylglyceride fatty acid
distribution of the same cultivar by gas chromatography (24)
showed linolic (44 area%), palmitic (25 area%), and linolenic
acids (20 area%) to be the main components, matching the data
reported by Tevini et al. concerning the total lipid fraction (15).
Interestingly, this distribution seems not to be reflected by the
native fatty acid pattern of carotenoid diesters but underlines
an overall tendency, previously found in other plants, such as
T. erecta (21). Further studies will have to clarify if there exists
an enzyme system that is responsible for the highly specific
esterification step in the plant carotenoid ester biosynthesis.
Figure 4. Mass spectrum (APCI+ mode) of violaxanthin myristate−
palmitate (corresponding to 10, Figure 3) as an example of the
fragmentation pattern of a mixed carotenoid diester possessing epoxy
groups. Fragments belonging to the main fragmentation pathway are
underlined.
LITERATURE CITED
(1) Iwanzik, W.; Tevini, M.; Stute, R.; Hilbert, R. Carotinoidgehalt
und -zusammensetzung verschiedener deutscher Kartoffelsorten
und deren Bedeutung fu¨r die Fleischfarbe der Knolle. Potato
Res. 1983, 26, 149-162.
cedure, using lipases for hydrolysis of triacylglycerides, was
applied. To ensure nonspecific hydrolysis of triacylglycerides,
a mixture of two commercial microbial lipases was used. Both
lipases were proven not to cleave carotenoid esters in earlier
experiments (data not shown), ensuring carotenoid ester stability
during the cleanup procedure. The extract cleaned in this way
was subjected to LC(APCI)-MS analysis again and showed
the chromatogram depicted in Figure 3, pointing to an extremely
complex carotenoid ester spectrum. The MS and the UV data
of several peaks (9-16) are summarized in Table 2. An
exemplary mass spectrum, obtained from a mixed xanthophyll
diester, is given in Figure 4; it represents violaxanthin
myristate-palmitate (peak 10, Figure 3). The fragmentation
pattern is charaterized by two pathways, showing daughter ions
with diverse intensity; signals with high intensity belong to the
main pathway (the respective fragment masses are underlined
in Figure 4), which is characterized by the following cascade:
m/z 1031 is generated by the loss of water from the quasimo-
lecular ion (m/z 1049 [M + H]+); two daughter ions (m/z 803
and 775) from m/z 1031 are formed by the loss of myristic (228
Da) or palmitic (256 Da) acids, respectively; m/z 547 is
generated by release of the remaining fatty acid. A minor
fragmentation pathway starts likewise from the quasimolecular
ion at m/z 1049 and leads to the corresponding fragment ions
m/z 821 (loss of C14:0), 793 (loss of C16:0), and 565, which
represents the backbone of violaxanthin. A comparable frag-
mentation scheme is observed for 9-13, allowing for identifica-
tion of the nature of the fatty acids bound to the parent
carotenoid. Of course, in the case of lutein and zeaxanthin
diesters (14-16), the fragmentation pathway initiated by the
loss of water from [M + H]+ does not appear. The assignment
of the parent carotenoid is sometimes difficult, but making use
of the absorption spectra obtained by diode array detection led
to the unequivocal identification of 9, 10, 14, 15, and 16 (see
Table 2). Concerning 11-13, the UV/vis spectra did not exactly
match the spectroscopic data of the major carotenoids mentioned
above, particularly not the spectrum of antheraxanthin, one of
the major carotenoids. Although the parent carotenoids of 11-
13 were not unequivocally identified, the cleanup procedure
proved to be highly efficient, allowing for the LC(APCI)-MS-
based identification of fatty acids acylating carotenoids, which
occur in minute amounts in plant extracts.
(2) Olmedilla, B.; Granado, F.; Blanco, I.; Vaquero, M.; Cajigal,
C. Lutein in patients with cataracts and age-related macular
degeneration: A long-term supplementation study. J. Sci. Food
Agric. 2001, 81, 904-909.
(3) Landrum, J. T.; Bone, R. A. Lutein, zeaxanthin, and the macular
pigment. Arch. Biochem. Biophys. 2001, 385, 28-40.
(4) Lu, W.; Haynes, K.; Wiley, E.; Clevidence, B. Carotenoid content
and color in diploid potatoes. J. Am. Soc. Hortic. Sci. 2001, 126,
722-726.
(5) Barua, A. B.; Olson, J. A. Xanthophyll epoxides, unlike
â-carotene monoepoxides, are not detectibly absorbed by hu-
mans. J. Nutr. 2001, 131, 3212-3215.
(6) Motoyuki, M. Production of colorful potato. Breeding of potato
cultivars containing anthocyanins or carotenoids. Techno News
1998, 66, 19-22.
(7) Granado, F.; Olmedilla, B.; Blanco, I.; Rojas-Hidalgo, E. Major
fruit and vegetable contributors to the main serum carotenoids
in the Spanish diet. Eur. J. Clin. Nutr. 1996, 50, 246-250.
(8) Kasim, M. Beitra¨ge zur qualitativen Analyse der Carotinoide
von Kartoffeln. Nahrung 1967, 11, 405-409.
(9) Kasim, M. Untersuchungen u¨ber die Zusammensetzung der
Kartoffel-Carotinoide. Nahrung 1967, 11, 411-415.
(10) Pendlington, S.; Dupont, M. S.; Trussell, F. J. The carotenoid
composition of Solanum tuberosum. Biochem. J. 1965, 94, 25p-
26p.
(11) Mu¨ller, H. Determination of the carotenoid content in selected
vegetables and fruit by HPLC and photodiode array detection.
Z. Lebensm. Unters. Forsch. A 1997, 204, 88-94.
(12) Stute, R.; Hilbert, R.; Iwanzik, W.; Tevini, M. Carotinoidgehalte
und -zusammensetzung verschiedener deutscher Kartoffelsorten
und deren Bedeutung fu¨r die Fleischfarbe der Knolle. Vero¨ff.
Arbeitsgem. Kartoffelforsch. 1982, 4, 15-22.
(13) Fishwick, M. J.; Wright, A. Isolation and characterization of
amyloplast envelope membranes from Solanum tuberosum.
Phytochemistry 1980, 19, 55-59.
(14) Tevini, M.; Iwanzik, W.; Scho¨necker, G. Analyse, Vorkommen
und Verhalten von Carotinoiden in Kartoffeln und Kartoffel-
produkten. Jahrbuch des Forschungskreises der Erna¨hrungsin-
dustrie e.V.; Forschungskreis der Erna¨hrungsindustrie eV: Bonn,
1984; pp 36-53.
(15) Tevini, M.; Scho¨necker, G.; Iwanzik, W.; Riedmann, M.; Stute,
R. Analyse, Vorkommen und Verhalten von Carotinoidestern
in Kartoffeln. Vero¨ff. Arbeitsgem. Kartoffelforsch. 1983, 5, 42-
47.
In contrast to previous results (14, 15), we detected in a native
extract from the cultivar Nicola myristic and palmitic acids as
the main fatty acids bound to carotenoids. It is apparent that
only several major carotenoid esters of one single cultivar were
(16) Brown, C. R.; Edwards, C. G.; Yang, C.-P.; Dean, B. B. Orange
flesh trait in potato: Inheritance and carotenoid content. J. Am.
Soc. Hortic. Sci. 1993, 118, 145-150.

Carotenoids in Potatoes
J. Agric. Food Chem., Vol. 50, No. 24, 2002 7181
(17) K’osambo, L. M.; Carey, E. E.; Misra, A. K.; Wilkes, J.;
Hagenimana, V. Influence of age, farming site, and boiling on
pro-vitamin A content in sweet potato (Ipomoea batatas (L.)
Lam.) storage roots. J. Food Compos. Anal. 1998, 11, 305-
321.
(23) Horwitz, W., private communication with the AOAC, November
2001.
(24) Breithaupt, D. E.; Schwack, W. Determination of free and bound
carotenoids in paprika (Capsicum annuum L.) by LC/MS. Eur
Food Res. Technol. 2000, 211, 52-55.
(18) Breithaupt, D. E. Enzymatic hydrolysis of carotenoid fatty acid
esters of red pepper (Capsicum annuum L.) by a lipase from
Candida rugosa. Z. Naturforsch. 2000, 55c, 971-976.
(19) Barua, A. B. Improved normal-phase and reversed-phase gradient
high-performance liquid chromatography procedures for the
analysis of retinoids and carotenoids in human serum, plant and
animal tissues. J. Chromatogr. A 2001, 936, 71-82.
(20) Barua, A. B. Intestinal absorption of epoxy-â-carotenes by
humans. Biochem. J. 1999, 339, 359-362.
(21) Breithaupt, D. E.; Wirt, U.; Bamedi, A. Differentiation between
lutein monoester regioisomers and detection of lutein diesters
from marigold flowers (Tagetes erecta L.) and several fruits by
liquid chromatography-mass spectrometry. J. Agric. Food
Chem. 2002, 50, 66-70.
(25) Ko¨st, H.-P.; Zweig, G.; Sherma, J. S. Plant Pigments. CRC
Handbook of Chromatography; CRC Press: Boca Raton, FL,
1988; Vol. I.
(26) Bilger, W.; Fisahn, J.; Brummet, W.; Kossmann, J.; Willmitzer,
L. Violaxanthin cycle pigment contents in potato and tobacco
plants with genetically reduced photosynthetic capacity. Plant
Physiol. 1995, 108, 1479-1486.
(27) M´ınguez-Mosquera, M.; Hornero-Me´ndez, D. Separation and
quantification of the carotenoid pigments in red peppers (Cap-
sicum annuum L.), paprika, and oleoresin by reversed-phase
HPLC. J. Agric. Food Chem. 1993, 41, 1616-1620.
Received for review July 8, 2002. Revised manuscript received
September 12, 2002. Accepted September 12, 2002.
(22) Official Methods of Analysis of the Association of Official
Analytical Chemists, 16th ed.; AOAC International: Gaithers-
burg, MD, 1996; Recovery studies, XViii, item 26.
JF0257953