Identification and quantification of astaxanthin esters in shrimp (Pandalus, borealis) and in a microalga (Haematococcus pluvialis) by liquid chromatography-mass spectrometry using negative ion atmospheric pressure chemical ionization

Article abstract of DOI:10.1021/jf049780b

Negative ion liquid chromatography-atmospheric pressure chemical ionization mass spectrometry [negative ion LC-(APCI)MS] was used for the identification of astaxanthin esters in extracts of commercial shrimp (Pandalus borealis) and dried microalga (Haematococcus pluvialis) samples. A cleanup step using a normal phase solid phase extraction (SPE) cartridge was applied prior to analysis. Recovery experiments with astaxanthin oleate as model compound proved the applicability of this step (98.5 ± 7.6%; n = 4). The assignment of astaxanthin esters in negative ion LC-(APCI)MS was based on the detection of the molecular ion (M.-) and the formation of characteristic fragment ions, resulting from the loss of one or two fatty acids. Quantification of individual astaxanthin esters was performed using an astaxanthin calibration curve, which was found to be linear over the required range (1-51 μmol/L; r² = 0.9996). Detection limits, based on the intensity of M.-, a signal-to-noise ratio of 3:1, and an injection volume of 20 μL, were estimated to be 0.05 μg/mL (free astaxanthin), 0.28 μg/mL (astaxanthin-C16:0), and 0.78 μg/mL (astaxanthin-C16:0/C16:0), respectively. This LC-(APCI)MS method allows for the first time the characterization of native astaxanthin esters in P. borealis and H. pluvialis without using time-consuming isolation steps with subsequent gas chromatographic analyses of fatty acid methyl esters. The results suggest that the pattern of astaxanthin-bound polyunsaturated fatty acids of P. borealis does not reflect the respective fatty acid pattern found in triacylglycerides. Application of the presented LC-(APCI)MS technique in common astaxanthin ester analysis will forestall erroneous xanthophyll ester assignment in natural sources.

Full text of DOI:10.1021/jf049780b

3870 J. Agric. Food Chem. 2004, 52, 3870−3875
Identification and Quantification of Astaxanthin Esters in
Shrimp (Pandalus borealis) and in a Microalga (Haematococcus
pluvialis) by Liquid Chromatography−Mass Spectrometry Using
Negative Ion Atmospheric Pressure Chemical Ionization
†
DIETMAR E. BREITHAUPT
Institut fu¨r Lebensmittelchemie, Universita¨t Hohenheim, Garbenstrasse 28,
D-70599 Stuttgart, Germany
Negative ion liquid chromatography-atmospheric pressure chemical ionization mass spectrometry
[negative ion LC-(APCI)MS] was used for the identification of astaxanthin esters in extracts of
commercial shrimp (Pandalus borealis) and dried microalga (Haematococcus pluvialis) samples. A
cleanup step using a normal phase solid phase extraction (SPE) cartridge was applied prior to analysis.
Recovery experiments with astaxanthin oleate as model compound proved the applicability of this
step (98.5 ( 7.6%; n ) 4). The assignment of astaxanthin esters in negative ion LC-(APCI)MS was
based on the detection of the molecular ion (M•-) and the formation of characteristic fragment ions,
resulting from the loss of one or two fatty acids. Quantification of individual astaxanthin esters was
performed using an astaxanthin calibration curve, which was found to be linear over the required
range (1-51 µmol/L; r 2 ) 0.9996). Detection limits, based on the intensity of M•-, a signal-to-noise
ratio of 3:1, and an injection volume of 20 µL, were estimated to be 0.05 µg/mL (free astaxanthin),
0.28 µg/mL (astaxanthin-C16:0), and 0.78 µg/mL (astaxanthin-C16:0/C16:0), respectively. This LC-
(APCI)MS method allows for the first time the characterization of native astaxanthin esters in P. borealis
and H. pluvialis without using time-consuming isolation steps with subsequent gas chromatographic
analyses of fatty acid methyl esters. The results suggest that the pattern of astaxanthin-bound
polyunsaturated fatty acids of P. borealis does not reflect the respective fatty acid pattern found in
triacylglycerides. Application of the presented LC-(APCI)MS technique in common astaxanthin ester
analysis will forestall erroneous xanthophyll ester assignment in natural sources.
KEYWORDS: LC-(APCI)MS; astaxanthin esters; shrimp; Pandalus borealis; Haematococcus pluvialis;
carotenoid
INTRODUCTION
(Figure 1), depending on the respective organism and the
storage site (e.g., flesh, skin, ovaries) (4).
Astaxanthin, a symmetric ketocarotenoid (3,3′-dihydroxy-â,â-
carotene-4,4′-dione), naturally occurs in a wide variety of marine
and aquatic organisms. The bright red to pink color of
Crustaceae (shrimp, krill) and Salmonidae (salmon, rainbow
trout) results from accumulation of astaxanthin (1). Because
animals are not capable of de novo synthesis of astaxanthin,
microalgae serve as their natural dietary source. Apart from
coloration, astaxanthin is considered to be essential for growth
processes and as a vitamin A precursor in fish (2, 3). Thus,
astaxanthin is commonly used as a feed supplement in world-
wide fish farming to grow healthy and well-colored individuals.
The astaxanthin molecule bears two hydroxy functions located
at C3/C3′ of the â-ionone moieties. Thus, astaxanthin may be
present in its free as well as in its mono- or diesterified form
A natural source of astaxanthin is the green microalga
Haematococcus sp., which accumulates high concentrations of
monoesterified astaxanthin (5). Nonesterified astaxanthin is
found in the yeast Phaffia rhodozyma (Xanthophyllomyces
dendrorhous), another high-accumulating organism (6). Analysis
of the highly complex ester pattern found in Haematococcus
sp. has attracted interest in the past. Yuan et al. (7, 8) provided
a high-performance liquid chromatography (HPLC) method for
the separation and analysis of astaxanthin esters present in H.
lacustris, but they did not focus on individual ester identification.
However, the authors noted that astaxanthin monoesters were
the main components, accompanied by only trace amounts of
astaxanthin diesters. Renstrom and Liaaen-Jensen (9) identified
xanthophyll esters by means of thin-layer chromatography
followed by gas chromatographic (GC) detection of fatty acid
methyl esters. Predominant fatty acids in astaxanthin esters of
† Telephone (49) 711-4594094; fax (49) 711-4594096; e-mail breithau@
uni-hohenheim.de.
10.1021/jf049780b CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/19/2004

Identification of Astaxanthin Esters
J. Agric. Food Chem., Vol. 52, No. 12, 2004 3871
Figure 1. Structures of astaxanthin mono- (1) and diesters (2) (R ) saturated or unsaturated alkyl chains).
MATERIALS AND METHODS
H. pluVialis were C16:0 (7%), C18:0 (7%), C19:0 (6%), C20:0
(25%), and C18:1 (56%). The presence of oleic acid as the major
fatty acid found in the astaxanthin ester fraction of H. pluVialis
was affirmed by other studies (10, 11).
Chemicals. Light petroleum (boiling fraction 40-60 °C), methanol,
ethyl acetate, and acetone were purchased from Merck (Darmstadt,
Germany); tert-butyl methyl ether (TBME), n-hexane, pyridine (99.8%,
over molecular sieve), astaxanthin (98%), and acyl chlorides (98-99%
each; lauroyl, myristoyl, palmitoyl, stearoyl, oleoyl, and linoleoyl
chloride) were obtained from Sigma-Aldrich (Taufkirchen, Germany).
Filter paper (no. 610) was from Schleicher & Schuell (Dassel,
Germany). All solvents were distilled before use. High-purity water
was prepared with a Milli-Q 185 Plus water purification system
(Millipore, Eschborn, Germany). SPE cartridges for normal phase
chromatography (Sep-Pak; 910 mg florisil plus/cartridge) were pur-
chased from Waters GmbH, Eschborn, Germany.
Renstrom and Liaaen-Jensen (9) provided further information
on the astaxanthin ester pattern found in shrimp (Pandalus
borealis). The main fatty acids esterified with astaxanthin were
C12:0 (4-5%), C14:0 (2-3%), C16:0 (3-8%), C16:1 (9-
10%), C18:1 (17-44%), C22:1 (21-39%), C20:5 (2-5%), and
C22:6 (5-19%). Accordingly, oleic acid and C22:1 formed the
principal part. Polyunsaturated fatty acids (PUFAs) such as
C20:5 (eicosapentaenoic acid) and C22:6 (docosahexaenoic
acid), typical native constituents of various fish oils, also
occurred in significant amounts. In an older study, fatty acids
bound to astaxanthin derived from brown shrimp (Crangon
Vulgaris Fabr.) were investigated by GC-FID (12). Again,
C14:0, C16:0, C16:1, C18:0, and C18:1 were identified, but
quantitative data were not given. A recent investigation on the
red crab langostilla (Pleuroncodes planipes) provided detailed
information on the composition of fatty acids acylating astax-
anthin in three diester fractions and in one monoester fraction
(13). The main fatty acids found in the diester fractions were
C14:0 (5-9%), C16:0 (17-30%), C20:0 (0-23%), C16:1ω-7
(12-15%), and C18:1 (15-24%), whereas C20:5 (47%) and
C22:6 (11%) were predominant in the monoester fraction.
Recently, the fatty acid pattern of esterified astaxanthin of
Antarctic krill (Euphausia superba) was published (14). Mo-
lecular masses of the carotenoid esters were measured by field
desorption mass spectrometry. The fatty acid moieties were
additionally transmethylated and analyzed by GC. The authors
reported only five fatty acids in the mono- and diester frac-
tions: C12:0, C14:0, C16:0, C16:1 and C18:1. Surprisingly,
PUFAs were not found in the respective fractions.
Samples. Four shrimp samples (P. borealis), produced from different
manufacturers (sample I, without brine; samples II-IV, with brine),
were obtained from local supermarkets. The brine was removed and
the shrimp were dried with a dust-free cloth prior to homogenization.
Two dried H. pluVialis samples were kindly provided by the Microalgal
Biotechnology Laboratory (Ben-Gurion University of the Negev, Israel;
Prof. S. Boussiba) and by Cyanotech Corp. (Hawaii Ocean Science
Technology Park).
Preparation of Samples. Extraction. Shrimp (150 g) were minced
with an Ultra Turrax T 25 (Janke & Kunkel, Staufen, Germany) for 1
min. An aliquot of the homogenate (80 g) was suspended in methanol/
ethyl acetate/light petroleum (1:1:1 v/v/v; 100 mL) by mixing thor-
oughly by hand. The extract was separated from the tissue by employing
a Bu¨chner funnel covered with a filter paper under vacuum (500 mbar).
Direct homogenization of shrimp samples with the ternary solvent
mixture resulted in a slurry, which easily blocked the Bu¨chner funnel.
Thus, this method was not applicable. To allow for complete extraction,
additional portions of solvent were poured onto the filter cake and
vacuumed until the filter cake was completely colorless (6 × 100 mL).
The resulting extract was transferred into a separating funnel. To assist
phase separation, 4 mL of a saturated sodium chloride solution was
added. The water layer was discarded, and the organic layer was
transferred into an Erlenmeyer flask, dried with anhydrous sodium
sulfate, filtered, and evaporated to dryness. For cleanup, the residue
was dissolved in n-hexane (3 mL). H. pluVialis samples (150 mg) were
suspended in water (10 mL), mixed with an aliquot of a saturated
sodium chloride solution (4 mL), and extracted with methanol/ethyl
acetate/light petroleum (1:1:1 v/v/v; 20 mL each) in a separating funnel
until the extracts were colorless (3 × 20 mL). To ensure complete
extraction, sonification (30 s) was applied prior the last extraction step.
The upper organic phases were combined and treated as described
above.
The present study uses HPLC coupled to mass spectrometric
detection to obtain a deeper insight into the astaxanthin ester
pattern of H. pluVialis and P. borealis without time-consuming
isolation steps of individual esters and without the use of
transmethylation of fatty acids. For removal of accompanying
lipids, normal phase solid phase extraction (SPE) was applied
as a single cleanup step. Because reference substances were not
available commercially, astaxanthin mono- and diesters were
synthesized independently to facilitate peak assignment in
complex mixtures. This study did not focus on the stereoisomeric
profile of astaxanthin esters or on the presence of minor
xanthophylls (e.g., echinenone, canthaxanthin).
Cleanup Procedure. Five SPE columns were directly coupled in-
line by sticking them together and conditioned with n-hexane. The use
of five columns was necessary to provide sufficient sorbent for good
separation. Each entire extract was poured completely onto the first
column using a syringe. For removal of interfering compounds, the
whole column was washed with n-hexane (30 mL) and mixtures of

3872 J. Agric. Food Chem., Vol. 52, No. 12, 2004
Breithaupt
Table 1. Negative Ion LC-(APCI)-MS Data (CV 45 V) Used for
acetone in n-hexane [10, 20, 30, 40, and 50% (v/v); 30 mL each].
Elution of astaxanthin esters adsorbed between columns two and four
was performed with pure acetone (30 mL). Colored fractions were
combined, the solvent was evaporated (50 mbar, 30 °C), and the residue
was dissolved in TBME/methanol (1:1 v/v; 3 mL). After membrane
filtration (0.45 µm), samples were subjected to HPLC with diode array
detector (DAD) or liquid chromatography-mass spectrometry using
an atmospheric pressure chemical ionization interface [LC-(APCI)MS]
analyses. All procedures were performed under dim light. Samples were
analyzed immediately after cleanup.
Identification of Astaxanthin Esters in H. pluvialis and Shrimp (P.
borealis) Extractsa
m/z
•-
•
•
compound
M
[M − FA ]-
[M − FA ]-
1
2
A−C18:4 (1)*b
A−C18:3 (2)*
A−C18:2 (3)
A−C18:1 (4)
A−C16:0 (5)
A−C18:0 (6)
A−C20:0 (7)*
A−C18:1/C18:3 (8)*
A−C18:1/C18:2 (9)
A−C16:0/C18:2 (10)
A−C18:1/C18:1 (11)
A−C16:0/C16:0 (12)
A−C18:0/C18:1 (13)
A(14)
A−C12:0 (15)
A−C14:0 (16)
A−C20:1 (17)*
A−C12:0/C:12:0 (18)
A−C12:0/C14:0 (19)
A−C12:0/C18:1 (20)
A−C12:0/C16:0 (21)
A−C12:0/C20:1 (22)*
A−C12:0/C22:1 (23)*
A−C18:1/C20:1 (24)*
854.6 (100%)
856.6 (100%)
858.6 (100%)
860.6 (100%)
834.6 (100%)
862.6 (100%)
890.7 (100%)
1120.8 (100%)
1122.9 (100%)
1096.9 (100%)
1124.9 (100%)
1072.8 (100%)
1126.9 (100%)
596.4 (100%)
778.6 (24%)
578.4 (85%)
578.4 (90%)
578.4 (130%)
578.4 (75%)
578.4 (127%)
578.4 (114%)
578.4 (33%)
838.6 (32%)
840.6 (12%)
840.6 (10%)
842.6 (22%)
816.6 (18%)
842.6 (13%)
EValuation of the Cleanup Step. To determine possible losses of
astaxanthin esters during the cleanup procedure, recovery experiments
using synthesized astaxanthin oleate as reference compound were
performed. Aliquots (2 mL) of an astaxanthin oleate solution (c ) 75
µg/mL in n-hexane) were placed with a syringe directly on the SPE
column and eluted as described (final volume ) 2 mL). The concentra-
tion of this sample is referred to as c2. For preparation of a reference
solution, another aliquot (2 mL) was evaporated under vacuum and
redissolved in TBME/methanol (1:1 v/v; 2 mL). The concentration of
this reference solution is referred to as c1. Recoveries of astaxanthin
oleate were calculated as follows: % recovery ) c2 × 100/c1. The
following recovery was obtained (n ) 4): 98.5 ( 7.6%.
Quantification of Astaxanthin Esters by HPLC-DAD. The
concentration of an astaxanthin standard solution was determined
spectrophotometrically. Because the solubility of astaxanthin is rather
poor in most solvents, astaxanthin (∼20 mg) was dissolved in toluene
(50 mL), although no ꢀmol value was available. The concentration
calculated using the data set given for canthaxanthin (ꢀmol ) 118200;
480 nm) (15) was 51.2 µmol/L (30.6 mg/L). Because the influence of
the two hydroxy groups of astaxanthin on the spectroscopic behavior
is negligible, quantitative results are considered not to be influenced.
For calibration, an aliquot (40 mL) was evaporated to dryness,
redissolved in TBME/methanol (1:1 v/v; 40 mL), and further diluted
to the required concentrations (1.0-51.2 µmol/L). Aliquots were
subjected immediately to HPLC-DAD analysis. For quantification, the
calibration curve was created by plotting the peak area (mAU × s)
versus the concentration (µmol/L). The calibration graph was found to
be linear over the required range (1-51 µmol/L; r2 ) 0.9996).
Concentrations of astaxanthin esters were calculated according to the
respective molecular masses.
842.6 (26%)
842.6 (14%)
816.6 (25%)
844.6 (16%)
578.4 (100%)
578.4 (100%)
578.4 (100%)
760.5 (173%)
788.6 (74%)
842.6 (71%)
816.6 (71%)
870.6 (46%)
898.7 (43%)
870.6 (76%)
806.6 (30%)
888.7 (31%)
960.7 (100%)
988.8 (100%)
1042.8 (100%)
1016.8 (100%)
1070.8 (100%)
1098.9 (100%)
1152.9 (100%)
760.5 (63%)
760.5 (109%)
760.5 (49%)
760.5 (68%)
760.5 (52%)
842.6 (77%)
a FA, fatty Acid; A, astaxanthin. Masses of the fragment ions and the signal
intensities (in parentheses) are given. Assignment of 1−24 corresponds to peak
numbering in Figure 3. b The asterisk indicates that assignment is based on the
fragmentation pattern of LC-(APCI)MS analyses only.
Optimization of LC-(APCI)MS Parameters. To optimize conditions
for LC-(APCI)MS analyses, a standard solution containing synthesized
homogeneous astaxanthin diesters was used in preliminary experiments.
Because xanthophylls are well-known to form protonated quasimo-
lecular ions [M + H]+ (16), the solution was first analyzed in the
positive ion mode (CV 30 V). As expected, the quasimolecular ions as
well as the respective fragment ions formed by loss of one or two fatty
acids were generated. However, when cleaned shrimp extracts were
analyzed, an extremely high background noise hampered the interpreta-
tion of the fragmentation pattern. These interferences may originate
from residual triacylglycerides, also forming positive ions. Increasing
the energy imparted to the analytes from 30 to 60 V enhanced
xanthophyll ester fragmentation. Subsequently, the standard mixture
was analyzed using the negative ion mode (CV 30, 45, 60 V). Under
these conditions, negatively charged molecular ions (M•-) were formed,
whereas [M - H]- ions were present with low intensity. The same
pattern was found with free astaxanthin. The mass spectra obtained
from standard solutions of free astaxanthin showed the predicted
molecular ion at m/z 596.4 (100%) and a M - 1 signal with a relative
intensity of 16%.
Synthesis and Isolation of Astaxanthin Esters. Astaxanthin (1 mg)
was dissolved in dry pyridine (5 mL) and reacted with the respective
acyl chloride as described earlier for zeaxanthin esters (16). Mixed esters
were prepared by adding both acyl chlorides dropwise one after another.
The crude products were subjected to semipreparative HPLC on C30
material as described by Weller and Breithaupt (16). For elution,
isocratic solvent mixtures consisting of TBME and methanol (v/v) were
used (for peak numbering see Table 1): 50:50 (6, 9, 13), 40:60 (3, 4,
5, 15, 16), 30:70 (12, 18, 20, 21), 20:80 (10, 11, 19). Retention times
were between 5 min (15) and 29 min (11). After isolation, combined
fractions of multiple separations were evaporated to dryness, im-
mediately redissolved in TBME/methanol (1:1 v/v; 3 mL), and stored
at - 20 °C. All standard solutions were analyzed by LC-(APCI)MS.
HPLC and LC-(APCI)MS. The HPLC consisted of a modular
system HP1100 (Hewlett-Packard GmbH, Waldbronn, Germany) with
diode array detector (480 nm). A 250 × 4.6 mm i.d. YMC analytical
column (YMC Europe, Schermbeck, Germany) equipped with 5 µm
C30 reversed phase material including a 10 × 4.0 mm i.d. precolumn
was used (35 °C). The mobile phase consisted of mixtures of methanol/
TBME/water [81:15:4 v/v/v (A) and 6:90:4 v/v/v (B)]. The following
gradient was used (min/% A): 0/99; 39/44; 45/0; 50/99; 55/99. The
flow rate was 1 mL/min and the injection volume, 20 µL. LC-(APCI)-
MS was performed on an HP1100 modular HPLC system, coupled to
a Micromass (Manchester, U.K.) VG platform II quadrupole mass
spectrometer. The following MS parameters were used: APCI source,
150 °C; APCI probe, 400 °C; corona voltage, 3.7 kV; HV lens, 0.5
kV. Nitrogen was used as sheath (75 L/h) and drying gas (300 L/h).
Various cone voltages (CV; 30, 45, 60 V) were tested in the positive
and in the negative ionization mode. Further details were given by
Breithaupt et al. (17).
Detection Limit of Astaxanthin Esters [LC-(APCI)MS]. For estimat-
ing the detection limits of astaxanthin esters and astaxanthin, synthesized
astaxanthin-C16:0, astaxanthin-C16:0/C16:0, and free astaxanthin
were employed as model compounds in equal concentrations (stock
solutions of 10 µmol/L each; TBME/methanol, 1:1 v/v). Solutions were
further diluted with TBME/methanol (1:1 v/v). On the basis of the
intensity of the respective molecular ion (M•-), a signal-to-noise ratio
of 3:1, and an injection volume of 20 µL, the detection limits were
estimated to be as follows: free astaxanthin, 0.08 µmol/L (0.05 µg/
mL); astaxanthin-C16:0, 0.33 µmol/L (0.28 µg/mL); astaxanthin-
C16:0/C16:0, 0.73 µmol/L (0.78 µg/mL).
RESULTS AND DISCUSSION
Fragmentation of Astaxanthin Esters. An example of a
mass spectrum obtained from a mixed astaxanthin ester
(C12:0/C16:0) determined in a shrimp (P. borealis) extract is

Identification of Astaxanthin Esters
J. Agric. Food Chem., Vol. 52, No. 12, 2004 3873
Figure 2. Mass spectrum of astaxanthin−C12:0/C16:0 [negative ion LC-(APCI)MS; CV 45 V], obtained from a shrimp (P. borealis) extract (A) and from
•-
•
•
a standard solution (B). Assignment of the respective ions (m/z) is as follows: 1016.8 (M ); 1015.8 ([M − H]-); 816.6 ([M − C12:0]-); 760.5 ([M −
•
C16:0]-); 564.4; 560.4 ([M − C12:0 − C16:0]-).
shown in Figure 2A. For direct comparison, the mass spectrum
of the chemically synthesized compound is presented in Figure
2B. The formation of M•- molecular ions has been described
previously by van Breemen et al. (18), who analyzed free
xanthophylls in the negative mode. Liquid chromatography
coupled to particle beam MS has been used earlier in the
negative mode for analyses of â-carotene solutions (19). The
fragmentation pattern of astaxanthin esters was dominated by
loss of fatty acids as neutral molecules, resulting in the case of
mixed diesters in the formation of two [M• - FA]- ions.
Enhancing the CV from 30 to 45 and 60 V resulted in a decrease
of the molecular ion intensity and in an increase of the fragment
ion abundance. The ion representing the backbone of astaxanthin
(m/z 560.4) was present in all spectra but varied in intensity
and was accompanied by fragment ions with low intensity
(Figure 2). Additionally, a fragment ion at m/z 564.4 was
present, even forming the base peak in some experiments. The
origin of this ion is as yet unknown. To obtain the best ratio
among intensities of molecular and fragment ions, the CV was
set to 45 V in all subsequent experiments. The mass spectro-
metric data used for the identification of astaxanthin esters in
H. pluVialis and P. borealis extracts are presented in Table 1.
The reason for the low abundance of monoester molecular ions
in the case of 15-24 (P. borealis) is not yet clear. However,
identification was achievable in most cases. It must be empha-
sized that relative fragment ion intensities are given for cleaned
real samples, not for synthesized reference compounds. The
backbone masses m/z 578.4 (monoesters) and m/z 560.4
(diesters) were used to scan the total ion current (TIC) trace
for the existence of astaxanthin esters. In a subsequent step,
individual mass spectra corresponding to peaks with absorption
at 480 nm were produced. Because only mass differences
between quasimolecular and fragment ions were used for
assignment of acyl chains, LC-MS does not allow for double-
bond location. If this information is required, GC represents
the method of choice. Alternatively, LC-MSn experiments have
to be carried out.
time window between 15 and 25 min, whereas diesters typically
showed retention times of 25-35 min (Figure 3A). All UV-
vis spectra featured one broad maximum at 480-482 nm (DAD,
measured in the eluent), which is typical for the astaxanthin
skeleton (15). To unequivocally allocate identification based on
LC-(APCI)MS, astaxanthin mono- and diesters, presumed to
be constituents of the extract, were synthesized independently
and added to aliquots of the extract. Cochromatography was
consistent with assignment of the respective compounds.
Unfortunately, only a few acyl chlorides were available com-
mercially. In particular, fatty acids of the C18 family with more
than two double bonds or compounds with chain lengths over
18 carbons were not obtainable. Accordingly, three astaxanthin
monoester (1, 2, 7) and one diester (8) were identified only on
the basis of their fragmentation pattern. Structural assignment
of compounds marked with an asterisk in Table 1 is conse-
quently not confirmed by an independent method. Interpretation
of the mass spectra of the broad peak 9 suggested the presence
of a coeluting component, which was presumed to be astaxan-
thin-C22:0. Synthesis of this ester was not possible, forestalling
conclusive identification. However, the main monoesters of H.
pluVialis unambiguously contained the acyl chains C18:3 (2),
C18:2 (3), C18:1 (4), and C16:0 (5). Thus, the predominance
of oleic acid, stated in the literature (e.g., ref 9), was confirmed
by the results of this study.
Astaxanthin Esters in P. borealis. Figure 3B shows a typical
chromatogram of a shrimp extract. In contrast to H. pluVialis,
astaxanthin diesters clearly formed the main components. Free
astaxanthin (14) appeared as a minor constituent. Chromato-
graphic resolution in the rear part of the chromatogram was
not excellent, but identification of some astaxanthin esters was
possible: five astaxanthin monoesters (4, 5, 15-17) and eight
astaxanthin diesters (11, 18-24) were identified. Lack of acyl
chlorides impeded cochromatography of three esters comprising
C20:1 (17, 22, 24) or C22:1 (23). Coelution with other
components was observed for 22 and 23 because signals
consistent with the fragmentation pathways of astaxanthin-
C14:0/C18:1 and astaxanthin-C14:0/C20:1, respectively, were
found. Although it was not possible to calculate the complete
astaxanthin ester pattern of P. borealis, lauric acid was found
Astaxanthin Esters in H. pluWialis. Astaxanthin monoesters,
the main derivatives of astaxanthin in H. pluVialis, eluted in a

3874 J. Agric. Food Chem., Vol. 52, No. 12, 2004
Breithaupt
Figure 3. Chromatograms (DAD, 480 nm, extended sections) of an H. pluvialis extract (A) and of a shrimp extract (P. borealis) (B). For peak assignment
see Table 1.
Table 2. Concentrations of Astaxanthin Esters (Micrograms per 100 g
of Dry Matter ± SD; n ) 4) in H. pluvialisa
Table 3. Concentrations of Astaxanthin Esters (Micrograms per 100 g
of Wet Matter ± SD; n ) 4) in Shrimp (P. borealis)a
compound
A−C18:4 (1)
A−C18:3 (2)
A−C18:2 (3)
A−C18:1 (4)
A−C16:0 (5)
A−C18:0 (6)
A−C20:0 (7)
A−C18:1/C18:3 (8)
A−C18:1/C18:2 (9)
A−C16:0/C18:2 (10)
A−C18:1/C18:1 (11)
A−C16:0/C16:0 (12)
A−C18:0/C18:1 (13)
sample I
sample II
compound
A(14)
A−C12:0 (15)
A−C14:0 (16)
A−C18:1 (4)
A−C16:0 (5)
A−C20:1 (17)
sample I
sample II
sample III
sample IV
26.8 ± 2.7
179.3 ± 13.4
372.7 ± 30.6
525.2 ± 37.7
182.6 ± 11.3
123.2 ± 9.5
63.9 ± 5.2
16.8 ± 1.0
87.6 ± 8.0
17.8 ± 2.2
36.3 ± 2.6
17.5 ± 2.8
18.6 ± 1.4
34.2 ± 0.6
76.8 ± 1.8
135.2 ± 2.8
173.0 ± 5.7
139.5 ± 4.2
32.5 ± 0.8
49.7 ± 1.3
12.2 ± 0.3
47.7 ± 1.1
23.0 ± 0.4
17.0 ± 0.3
16.7 ± 0.3
6.5 ± 0.3
10.0 ± 0.8
48.0 ± 3.2
5.7 ± 0.9
25.7 ± 1.5
5.7 ± 0.3
6.6 ± 0.5
14.1 ± 1.8
53.0 ± 2.5
5.7 ± 0.3
28.3 ± 1.6
6.5 ± 0.2
6.4 ± 0.2
18.0 ± 2.6
59.5 ± 2.1
4.9 ± 0.3
26.2 ± 0.8
9.2 ± 0.3
4.1 ± 0.4
5.4 ± 0.6
44.7 ± 1.8
3.4 ± 0.1
15.4 ± 0.3
5.6 ± 0.1
3.6 ± 0.2
A−C12:0/C:12:0 (18) 129.1 ± 4.8 146.2 ± 7.4 116.2 ± 3.2 151.1 ± 3.8
A−C12:0/C14:0 (19)
A−C12:0/C18:1 (20)
A−C12:0/C16:0 (21)
A−C12:0/C20:1 (22)
A−C18:1/C18:1 (11)
A−C12:0/C22:1 (23)
A−C18:1/C20:1 (24)
22.7 ± 1.5
92.3 ± 5.3 100.6 ± 2.9
23.6 ± 1.7
9.0 ± 0.4
81.1 ± 2.4
33.2 ± 1.1
12.7 ± 1.1
24.6 ± 0.8
15.5 ± 0.5
6.7 ± 0.5
14.4 ± 0.7
77.2 ± 1.8
29.5 ± 0.7
23.8 ± 0.8
17.9 ± 0.4
12.6 ± 0.3
6.9 ± 0.3
24.2 ± 1.4
38.8 ± 2.0
17.7 ± 1.5
16.8 ± 1.0
12.0 ± 0.8
28.6 ± 0.8
37.9 ± 2.7
23.0 ± 0.7
19.8 ± 1.4
13.2 ± 0.4
a A, astaxanthin. Quantitative amounts were calculated on the basis of the
respective molecular masses. Numbers of the compounds correspond to Figure
3A.
a A, astaxanthin. Quantitative amounts were calculated based on the respective
molecular mass. Numbers of the compounds correspond to Figure 3B.
to be one of the prevalent fatty acids. PUFAs, formerly regarded
as typical components in the astaxanthin ester fraction of P.
borealis (9), were not found as main fatty acids in this study.
Thus, the distribution of fatty acids bound to astaxanthin seems
not to reflect the native pattern of PUFAs found in the
triacylglyceride fraction (20). However, this observation has to
be proven in further studies using GC analyses of methylated
fatty acids and the method presented here. Discrepancies in the
fatty acid pattern of the triacylglyceride fraction on the one hand
and the xanthophyll ester fraction on the other hand were already
observed by investigating the lutein ester pattern of Tagetes
erecta (17) and seem to be also valid for P. borealis.
Quantification of Astaxanthin Esters. To determine the
concentrations of individual astaxanthin esters by HPLC-DAD,
free astaxanthin was used to create a calibration curve (mAU
× s vs µmol/L). Assuming similar extinction coefficients of
free and acylated forms, this method allows for quantification
of individual esters, considering the respective molecular masses.
Quantitative results are presented in Tables 2 (H. pluVialis) and
3 (P. borealis). In both H. pluVialis samples, astaxanthin-C18:1
(4) was the main astaxanthin ester; however, concentrations
varied considerably between the two samples (525 vs 173 µg/
100 g). This may be due to unknown growth and/or storage
conditions. Quantitative data of P. borealis proved the homo-
geneous diester astaxanthin-C12:0/C12:0 (18) to be the pre-
dominant compound, followed by a mixed diester astaxanthin-
C12:0/C18:1 (20). Notably, the results of samples II-IV, which
contained brine, did not differ from those obtained for sample
I, containing no brine. Although preliminary saponification
experiments using methanolic potassium hydroxide solution led
to the disappearance of the respective peaks in H. pluVialis and
P. borealis extracts, total free astaxanthin was not quantified
because special care has to be taken to avoid byproduct
formation (21).
Taken together, negative ion LC-(APCI)MS provides a
possibility to analyze astaxanthin esters in extracts of H.
pluVialis and P. borealis after SPE cleanup without time-
consuming isolation or derivatization steps. Quantification can
be performed using astaxanthin as reference material. Applica-
tion of the presented LC-(APCI)MS technique in common
astaxanthin ester analysis will forestall erroneous xanthophyll
ester assignment in natural sources in future experiments.
ACKNOWLEDGMENT
Thanks go to Professor Dr. W. Schwack, University of Hohen-
heim, for the excellent working conditions at the Institute of

Identification of Astaxanthin Esters
J. Agric. Food Chem., Vol. 52, No. 12, 2004 3875
(12) Snauwaert, F.; Tobback, P. P.; Maes, E. Carotenoid pigments
in the brown shrimp (Crangon Vulgaris Fabr.). Lebensm. Wiss.
Technol. 1973, 6, 43-47.
Food Chemistry. Special thanks are due to Anne Klo¨ppel for
the ambitious work during her research course in food chemistry.
(13) Coral-Hinostroza, G. N.; Bjerkeng, B. Astaxanthin from the red
crab langostilla (Pleuroncodes planipes): Optical R/S isomers
and fatty acid moieties of astaxanthin esters. Comp. Biochem.
Physiol. B 2002, 133, 437-444.
(14) Takaichi, S.; Matsui, K.; Nakamura, M.; Muramatsu, M.; Hanada,
S. Fatty acids of astaxanthin esters in krill determined by mild
mass spectrometry. Comp. Biochem. Physiol. B 2003, 136, 317-
322.
(15) Ko¨st, H.-P., Ed. CRC Handbook of Chromatography, Plant
Pigments; CRC Press: Boca Raton, FL, 1988; Vol. 1, p 129.
(16) Weller, P.; Breithaupt, D. E. Identification and quantification
of zeaxanthin esters in plants using liquid chromatography-mass
spectrometry. J. Agric. Food Chem. 2003, 51, 7044-7049.
(17) 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.
(18) van Breemen, R. B.; Huang, C. R.; Tan, Y.; Sander, L. S.;
Schilling, A. B. Liquid chromatography/mass spectrometry of
carotenoids using atmospheric pressure chemical ionization. J.
Mass Spectrom. 1996, 31, 975-981.
(19) Careri, M.; Manini, P. Liquid chromatography-particle-beam
mass spectrometric determination of â-carotene by negative-ion
chemical ionization. Chromatographia 1997, 46, 425-429.
(20) Guillou, A.; Khall, M.; Adambounou, L. Effects of silage
preservation on astaxanthin forms and fatty acid profiles of
processed shrimp (Pandalus borealis) waste. Aquaculture 1995,
130, 351-360.
LITERATURE CITED
(1) Britton, G., Liaaen-Jensen, S., Pfander, H., Eds. Carotenoids Vol.
1A: Isolation and Analysis; Birkha¨user Verlag: Basel, Swit-
zerland, 1999.
(2) Torissen, O. J.; Hardy, R. W.; Shearer, K. D. Pigmentation of
salmonidsscarotenoid deposition and metabolism. ReV. Aquat.
Sci. 1989, 1, 209-225.
(3) Gobantes, I.; Choubert, G.; Milicua, J. C. G.; Go´mez, R. Serum
carotenoid concentration changes during sexual maturation in
farmed rainbow trout (Oncorhynchus mykiss). J. Agric. Food
Chem. 1998, 46, 383-387.
(4) Schiedt, K.; Leuenberger, F. J.; Vecchi, M.; Glinz, E. Absorption,
retention and metabolic transformation of carotenoids in rainbow
trout, salmon and chicken. Pure Appl. Chem. 1985, 57, 685-
692.
(5) Johnson, E. A.; An, G. H. Astaxanthin from microbial sources.
Crit. ReV. Biotechnol. 1991, 11, 297-326.
(6) Parajo, J. C.; Santos, V.; Vazquez, M. Production of carotenoids
by Phaffia rhodozyma growing on media made from hemicel-
lulosic hydrolysates of Eucalyptus globulus wood. Biotechnol.
Bioeng. 1998, 59, 501-506.
(7) Yuan, J.-P.; Gong, X.-D.; Chen, F. Separation and analysis of
carotenoids and chlorophylls in Haematococcus lacustris by
high-performance liquid chromatography photodiode array detec-
tion. J. Agric. Food Chem. 1997, 45, 1952-1956.
(8) Yuan, J.-P.; Gong, X.-D.; Chen, F. Separation and identification
of astaxanthin esters and chlorophylls in Haematococcus lacustris
by HPLC. Biotechnol. Tech. 1996, 10, 655-660.
(9) Renstrom, B.; Liaaen-Jensen, S. Fatty acid composition of some
esterified carotenols. Comp. Biochem. Physiol. 1981, 69B, 625-
627.
(10) Kobayashi, M.; Kakizono, T.; Nagai, S. Astaxanthin production
by a green alga, Haematococcus pluVialis accompanied with
morphological changes in acetate media. J. Ferment. Bioeng.
1991, 71, 335-339.
(21) Yuan, J.-P.; Chen, F. Hydrolysis kinetics of astaxanthin esters
and stability of astaxanthin of Haematococcus pluVialis during
saponification. J. Agric. Food Chem. 1999, 47, 31-35.
(11) Grung, M.; D’Souza, F. M. L.; Borowitzka, M.; Liaaen-Jensen,
S. Algal carotenoids. 51. Secondary carotenoids 2. Haemato-
coccus pluVialis aplanospores as a source of (3S,3′S)-astaxanthin
esters. J. Appl. Phycol. 1992, 4, 165-171.
Received for review February 9, 2004. Revised manuscript received
April 14, 2004. Accepted April 14, 2004.
JF049780B