2358 J. Agric. Food Chem., Vol. 55, No. 6, 2007
Matsumoto et al.
samples were immediately frozen in liquid nitrogen and kept at -80
C until use. Under these conditions, the carotenoids observed in this
study were stable in the samples for at least 2 years.
YMC, Kyoto, Japan) and ternary gradient elution with water (eluent
A), MeOH (eluent B), and MTBE (eluent C) as mobile phases at a
flow rate of 1 mL/min. The column temperature was kept at 25 °C.
The linear gradient program was performed as follows: the initial
condition was 5% A/90% B/5% C; 0-12 min, 95% B/5% C; 12-20
min, 86% B/14% C; 20-30 min, 75% B/25% C; 30-50 min, 50%
B/50% C; 50-60 min, 37.5% B/62.5% C; 60-73 min, 5% B/95% C.
LC-MS Conditions. Analyses were performed on an Agilent 1100
series HPLC system (Agilent Technologies, Palo Alto, CA) consisting
of a quaternary pump, an online degasser, a photodiode array detector
(set to scan 220-550 nm), and a temperature-controlled autosampler
(set at 4 °C) coupled to a PE SCIEX API2000 triple-stage quadrupole
tandem mass spectrometer (Applied Biosystems, Foster City, CA) with
a heated nebulizer ionization source. The analytes were detected using
atmospheric pressure chemical ionization (APCI) in a positive mode.
The curtain gas (nitrogen), nebulizer gas (air), and auxiliary gas (air)
were used at pressures of 45, 60, and 20 psi, respectively. The ion
source temperature was maintained at 400 °C. The nebulizer current
was set at 1 µA. The quantification of carotenoids was performed in
the selected ion monitoring (SIM) mode with a dwell time of 77 ms.
The Q-pole parameter (DP, EP, FP) was optimized for each carotenoid.
Data collection and peak integration were performed using Analyst
software (version 1.4, Applied Biosystems).
Quantification and Method Validation. To correct the response
variation in the mass spectrometer, the internal standard, citranaxanthin,
was added into all standard solutions and citrus extracts just before
the solutions were injected to LC-MS. The final concentration of the
internal standard was 0.14 µg/mL. Calibration curves were constructed
using standard solutions of the carotenoids at five different concentra-
tions in MTBE/MeOH (1:1, v/v). The linearity of the calibration curve
was confirmed by plotting the peak area ratio of the analyte to the
internal standard versus the concentration. The intraday and interday
precisions were calculated by replicate analyses (n ) 6) of the same
sample and expressed as the coefficient of variation (CV %). The accu-
racy of the method was calculated by (observed concentration/known
concentration) × 100. The recovery of each carotenoid during the
extraction was determined according to the ratio of the amount added
to that measured experimentally after the extraction. The assessment
of the matrix effect derived from the difference in the varieties in caro-
tenoid analysis was carried out according to the method of Matuszewski
et al. (28). Matrix-matched calibration curves were constructed using
the flavedo and juice sacs of five varieties. By comparing the slopes
of the calibration curves, the absence and presence of a matrix effect
were assessed. The lower limit of quantification (LLOQ) for each
analyte was determined with a signal-to-noise ratio of 10:1.
Cluster Analysis. Hierarchical clustering was performed by Ward’s
method in the application software JMP 6 (JMP release 6.0; SAS
Institute Inc., Cary, NC). First, to group the pattern of seasonal changes
in each carotenoid on the basis of their similarity, the carotenoid
concentrations from October to January of 39 varieties were standard-
ized and, subsequently, used for cluster analysis. In most carotenoids,
the cluster analysis classified the patterns of seasonal changes into three
groups, H, M, and L, in which carotenoid concentrations were high,
medium, and low, respectively. Second, to classify citrus varieties on
the basis of the patterns of seasonal changes of carotenoids, the
categories of the patterns of the seasonal changes, H, M, and L, were
converted into numerical values (ordinal scale), that is, 3, 2, and 1,
respectively. By cluster analysis, 39 varieties were classified on the
basis of similarity in the numerical values for the patterns of seasonal
changes in phytoene, â-cryptoxanthin, and violaxanthin. A previous
study (12) showed that two carotenoids, cis-violaxanthin and â-cryp-
toxanthin, were strong determinants in citrus classification. In the
present study, the concentrations of not only the two carotenoids but
also phytoene were higher than those of others, and the difference in
the concentrations among the varieties was larger than that in other
carotenoids. Thus, three carotenoids, phytoene, â-cryptoxanthin, and
violaxanthin, were selected to classify the citrus varieties.
°
Chemicals. all-trans-R-Carotene (â,ꢀ-carotene), all-trans-â-carotene
â,â-carotene), and all-trans-lycopene (ψ,ψ-carotene) were obtained
(
from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). all-trans-
Antheraxanthin (5,6-epoxy-5,6-dihydro-â,â-carotene-3,3′-diol) and all-
trans-neoxanthin (5′,6′-epoxy-6,7-didehydro-5,6,5′,6′-tetrahydro-â,â-
carotene-3,5,3′-triol) were obtained from DHI Water and Environment
(
Horsholm, Denmark). all-trans-Zeaxanthin (â,â-carotene-3,3′-diol) was
obtained from Extrasynthese (Genay, France). all-trans-â-Cryptoxanthin
â,â-caroten-3-ol) was obtained from Sokenkagaku (Tokyo, Japan). all-
trans-Neurosporene (7,8-dihydro-ψ,ψ-carotene), all-trans-γ-carotene
â,ψ-carotene), all-trans-lutein epoxide (5,6-epoxy-5,6-dihydro-â,ꢀ-
(
(
carotene-3,3′-diol), all-trans-violaxanthin (5,6,5′,6′-diepoxy-5,6,5′,6′-
tetrahydro-â,â-carotene-3,3′-diol), and all-trans-mutatoxanthin (5,8-
epoxy-5,8-dihydro-â,â-carotene-3,3′-diol) were obtained from Carote-
Nature (Lupsingen, Switzerland). all-trans-Lutein (â,ꢀ-carotene-3,3′-
diol) was purchased from Fluka Chemie (Buchs, Switzerland). Phytoene
(7,8,11,12,7′,8′,11′,12′-octahydro-ψ,ψ-carotene), phytofluene (7,8,11,-
12,7′,8′-hexahydro-ψ,ψ-carotene), three isomers of ú-carotene (7,8,7′,8′-
tetrahydro-ψ,ψ-carotene), cis-violaxanthin (5,6,5′,6′-diepoxy-5,6,5′,6′-
tetrahydro-â,â-carotene-3,3′-diol), and cis-antheraxanthin were isolated
from citrus flavedo and tentatively identified according to the previously
reported methods (13, 21-26). These carotenoids were used as authentic
standards. â-Apo-8′-carotenal was purchased from Fluka Chemie.
Potassium hydroxide (KOH), 2,6-di-tert-butyl-4-methylphenol (BHT),
diethyl ether, HPLC grade methyl tert-butyl ether (MTBE), methanol
(MeOH), and other chemicals were purchased from Wako Pure
Chemical Ind., Ltd.
Preparation of the Internal Standard, Citranaxanthin. Citrana-
xanthin was synthesized and identified according to the procedures
described by Stewart et al. (27). Briefly, 0.5 mL of acetone and 1 mL
of 10% methanolic KOH were added into a solution of â-apo-8′-
carotenal (0.3 mmol in 10 mL of diethyl ether/methanol, 1:1) at 0 °C.
The mixture was stirred for 18 h in the dark at room temperature in an
argon atmosphere. The mixture was transferred to a separatory funnel
containing 100 mL of diethyl ether. The ether phase was washed five
times with NaCl-saturated water to remove KOH and treated with
2 4
anhydrous Na SO . After the ether phase was evaporated, the residue
was purified using silica gel column chromatography using chloroform/
MeOH (40:0.5, v/v). The dark red crystal of the product was obtained
(122 mg). The product was analyzed by high-resolution fast atom
bombardment mass spectrometry (HRMS-FAB) (Mstation-700 double-
focusing magnetic sector mass spectrometer, JMS-700, JEOL, Tokyo,
Japan) and UV-vis spectrophotometry (UV-2200 spectrophotometer,
Shimadzu, Kyoto, Japan) and identified as citranaxanthin: UV-vis
+
(
C
hexane) λmax, nm, 493, 464; HRMS-FAB (m/z) [M + H] calcd for
33
H
45O, 457.3470; found, 457.3452.
Carotenoid Extraction. Samples (ca. 3 g of flavedo and ca. 5 g of
juice sacs) were homogenized in 50 mL of 40% (v/v) aqueous methanol
containing 0.5 g of basic magnesium carbonate using a homogenizer
(
Polytron homogenizer T-25). After vacuum filtration, the pigment in
the residue was extracted with diethyl ether/MeOH (7:3, v/v) containing
.1% (w/v) BHT until the residue was colorless. The extract was
0
transferred to a separatory funnel containing 100 mL of diethyl ether.
The ether phase was washed two times with NaCl-saturated water and
treated with anhydrous Na SO . After the ether phase was evaporated,
2 4
the residue was dissolved with 10 mL of diethyl ether, mixed with 5
mL of 20% methanolic KOH, and placed for 18 h in the dark in an
argon atmosphere at room temperature. The alkaline mixture was
transferred to a separatory funnel containing 100 mL of diethyl ether.
The ether phase was washed five times with NaCl-saturated water and
2 4
treated with anhydrous Na SO . After the ether phase was evaporated,
the residue was transferred to a 5 mL volumetric flask in which the
volume was adjusted with MTBE/MeOH (1:1, v/v) containing 0.1%
BHT to 5 mL. The sample was filtered through a 0.2 µm filter
(
Millipore, Bedford, MA) and stored at -30 °C before injection into
RESULTS
the LC-MS. Five microliters of the solution was injected.
Chromatographic Conditions. The separation of carotenoids was
carried out using a C30 carotenoid column (250 × 4.6 mm i.d., 5 µm,
Identification of Carotenoids in Citrus. The identification
of the carotenoids in flavedo and juice sacs was carried out using