Marc David Grynbaum
`Petra Hentschel
`Karsten Putzbach
`Jens Rehbein
`Manfred Krucker
`Graeme Nicholson
`Klaus Albert
`
`Institute of Organic Chemistry,
`University of Tuebingen,
`Tuebingen, Germany
`
`Grynbaum, Hentschel, Putzbach, Rehbein, Krucker, Nicholson, Albert
`
`1685
`
`Unambiguous detection of astaxanthin and
`astaxanthin fatty acid esters in krill (Euphausia
`superba Dana)
`
`HPLC atmospheric pressure chemical ionization (APCI)/MS, GC MS, HPLC diode
`array detection (DAD), and NMR were used for the identification of astaxanthin and
`astaxanthin fatty acid esters in krill (Euphausia superba Dana). Matrix solid phase
`dispersion was applied for the extraction of the carotenoids. This gentle and expedi-
`tious extraction technique for solid and viscous samples leads to distinct higher
`enrichment rates than the conventional liquid–liquid extraction. The chromatographic
`separation was achieved employing a C30 RP column that allows the separation of
`shape-constrained geometrical isomers. A methanol/tert-butylmethyl ether/water gra-
`dient was applied. (all-E) Astaxanthin and the geometrical isomers were identified by
`HPLC APCI/MS, by coelution with isomerized authentical standard, by UV spectros-
`copy (DAD), and three isomers were unambiguously assigned by microcoil NMR
`spectroscopy. In this method, microcoils are transversally aligned to the magnetic
`field and have an increased sensitivity compared to the conventional double-saddle
`Helmholtz coils, thus enabling the measurement on small samples. The carotenol
`fatty acid esters were saponified enzymatically with Lipase type VII from Candida
`rugosa. The fatty acids were detected by GC MS after transesterification, but also
`without previous derivatization by HPLC APCI/MS. C14:0, C16:0, C16:1, C18:1,
`C20:0, C20:5, and C22:6 were found in astaxanthin monoesters and in astaxanthin
`diesters. (all-E) Astaxanthin was identified as the main isomer in six fatty acid ester
`fractions by NMR. Quantitation was carried out by the method of internal standard.
`(13-cis) Astaxanthin (70 lg/g), 542 lg/g (all-E) astaxanthin, 36 lg/g unidentified asta-
`xanthin isomer, 62 lg/g (9-cis) astaxanthin, and 7842 lg/g astaxanthin fatty acid
`esters were found.
`
`Key Words: Astaxanthin; Astaxanthin fatty acid esters; Euphausia superba Dana; Krill; HPLC
`APCI/MS;
`
`Received: April 6, 2005; revised: May 31, 2005; accepted: June 1, 2005
`
`DOI 10.1002/jssc.200500152
`
`OriginalPaper
`
`1 Introduction
`
`Carotenoids are one of the most important groups of nat-
`ural pigments occurring in plants and animals. The first
`publication on these pigments dates back to 1817 and
`dealt with red pepper [1]. Carotenoids occur in plants,
`algae, and photosynthetic bacteria, where they play a criti-
`cal role in the photosynthetic process. They also appear in
`some nonphotosynthetic bacteria, yeasts, and molds,
`where they may carry out a protective function against
`damage by light and oxygen. Animals seem to be incap-
`able of synthesizing carotenoids de novo; those carote-
`noids which are present are of dietary source. Carote-
`noids provide animals with bright coloration and serve as
`antioxidants and radical scavengers [2, 3].
`
`Correspondence: Professor Klaus Albert, Universität Tübingen,
`Institut für Organische Chemie, Auf der Morgenstelle 18, D-
`72076 Tübingen, Germany. Fax: +49-7071-29-5875.
`E-mail: klaus.albert@uni-tuebingen.de.
`
`Figure 1. Chemical structure and numbering of astaxanthin
`and astaxanthin fatty acid esters.
`
`In marine invertebrate animals, pigmentation is often due
`to astaxanthin (Fig. 1), a symmetric ketocarotenoid (3,39-
`dihydroxy-b,b9-carotin-4,49-dione). Crustaceans (shrimp,
`krill) and Salmonidae (salmon, rainbow trout) have bright
`red to pink coloring due to the accumulation of astaxanthin
`[4]. It is the most commonly used carotenoid in salmonid
`fish farming and is deposited in the supplied, unesterified
`
`J. Sep. Sci. 2005, 28, 1685 – 1693
`
`www.jss-journal.de
`
`i 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`RIMFROST EXHIBIT 1039 page 0001
`
`

`

`1686
`
`Grynbaum, Hentschel, Putzbach, Rehbein, Krucker, Nicholson, Albert
`
`form in fish muscle [5–7]. Natural sources of astaxanthin
`such as krill or algae supply it in the esterified form. After
`hydrolysis of the esters it is deposited in the flesh of sal-
`monids in the free form [6, 8].
`
`Krill (Euphausia superba) is the most important zooplank-
`ton species and plays a key role in the Antarctic food web.
`It is a shrimp-like crustacean, which attains a size of 6 cm
`and feeds primarily on phytoplankton or sea ice algae. Krill
`is the staple food of many fish, birds, and mammals in the
`Southern Ocean. The biomass of Antarctic krill is consid-
`ered to be larger than that of the earth’s human population
`and krill swarms can occupy an area of 450 km2. The krill
`has been the subject of many investigations. In 1977, a
`total content of 15.7 lg/g biomass of cryptoxanthin, asta-
`xanthin esters, hydroxy-f-carotene, dihydroxy-f-caro-
`tene, zeaxanthin,
`flavoxanthin, and astaxanthin was
`found by column chromatography and UV/VIS spectro-
`scopy [9]. The carotenoid content has been investigated
`as a function of age, structure, and sex [10]. Carotenoid
`concentrations of 50 lg/g (male) and of 152 lg/g (female)
`were found and additionally small amounts of b-carotene
`and an unidentified polyoxyxanthophyll were described.
`Yamaguchi et al. [11] indicated that the carotenoid content
`of krill and krill meals is composed almost exclusively of
`astaxanthin and its monoesters and diesters (30–40 lg/g;
`5–15% unidentified). The stability of carotenoid pigments
`has also been studied [12], and the diesters have been
`found to be the most stable amongst them. The greatest
`carotenoid concentration was found in the cephalothorax
`and the carapace (71.3 and 59.8 lg/g) [13]. Maoka et al.
`[14] separated the two enantiomers (3S,39S-/3R,39R-) and
`the meso-astaxanthin (3S,39R/3R,39S), finding 3R,39R- to
`be the most abundant (62–71%). They found up to
`908 lg/g in krill eyes. The relative composition of asta-
`xanthin, its monoesters and diesters remains constant at
`5, 49, and 40%, respectively, but the total amount of
`astaxanthin varies with the season and the stage of sexual
`maturity [15].
`
`The lipid, sterol, and fatty acid composition of Antarctic
`krill has been investigated separately from the carote-
`noids [16]. Phosphatidylcholine, phosphatidylethanola-
`mine, triacylglycerol, free fatty acids, and sterols have
`been identified. The major fatty acids were C14:0, C16:0,
`C16:1(n-7), C18:1(n-9), C18:1(n-7), C20:5(n-3), and
`C22:6(n-3).
`
`Recently, Takaichi et al. [17] determined the fatty acids
`present in astaxanthin esters by mild MS. The extract was
`separated by HPLC on a C18 column, but the peaks are
`poorly resolved on their chromatograms. The peaks were
`collected and subjected to field desorption MS. Only five
`fatty acids were detected in astaxanthin, namely, dode-
`canoate, tetradecanoate, hexadecanoate, hexadeceno-
`
`ate, and octadecenoate. Surprisingly, polyunsaturated
`fatty acids were not found in the respective fractions.
`
`The present study uses HPLC coupled to atmospheric
`pressure chemical ionization (APCI) to study the carote-
`noid and carotenol fatty acid ester composition of Antarc-
`tic krill Euphausia superba Dana. Matrix solid phase dis-
`persion (MSPD) was used for the extraction of the carote-
`noids, and C30 sorbent was used for the separation. The
`identification of the carotenoids and its fatty acid esters
`was achieved by HPLC APCI/MS and microcoil NMR. The
`major advantage of HPLC APCI/MS is the direct detection
`of the fatty acid esters without the need of derivatization or
`fraction collection beforehand.
`
`As carotenoids are very sensitive to both light and air, an
`optimized combination of analytical sample preparation,
`separation, and detection techniques has to be used.
`MSPD, a very gentle and expeditious extraction technique
`for solid and viscous samples, was used for the extraction
`of the krill’s carotenoids and its esters. It is convenient to
`work with, decreases solvent use by up to 98%, and
`reduces sample turnaround time by 90% compared to
`conventional extraction techniques like liquid–liquid
`extraction [18–20]. Another advantage of MSPD is the
`higher enrichment of the analytes, which is very important
`for successful NMR measurements.
`
`After MSPD, the carotenoids were separated on C30 sor-
`bent. Triacontyl phases have been especially developed
`for the separation of shape-constrained natural com-
`pounds, such as carotenoid stereoisomers, and the sam-
`ple loading capacity of these phases is superior to that of
`the conventionally employed C18 materials. Detailed NMR
`studies investigating the structural parameters of C30 and
`interactions responsible for its specific separation beha-
`vior have already been carried out [21–23]. Online HPLC
`APCI/MS enables coupling of MS and LC with flow rates
`up to 1 mL/min and has a higher sensitivity than ESI for
`weakly polar compounds like carotenoids. APCI/MS
`allows the unambiguous identification and assignment of
`different carotenoids in positive as well as in negative ioni-
`zation mode [24–26]. Carotenol fatty acid esters can also
`be detected [27, 28]. Obviously, a distinction between
`stereoisomers is not possible by MS. For unambiguous
`structure elucidation, NMR spectroscopy is absolutely
`necessary. The chemical shifts and coupling constants of
`the stereoisomers are slightly changed [20]. By the intro-
`duction of a cis double bond, the symmetry is lost and the
`“inner” protons of the cis bond are shifted to the deep field
`while the “outer” protons are shifted to higher field. Sole-
`noidal NMR probes overcome the limited sensitivity of
`conventional saddle-shaped NMR probes and their sensi-
`tivity is threefold higher [29, 30]. They have successfully
`been employed not only for online coupled capillary HPLC
`NMR measurements [30–32], but also for stopped-flow
`
`J. Sep. Sci. 2005, 28, 1685 – 1693
`
`www.jss-journal.de
`
`i 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`RIMFROST EXHIBIT 1039 page 0002
`
`

`

`Astaxanthin and Astaxanthin Fatty Acid Esters in Krill
`
`1687
`
`measurements [33]. Mass-limited samples can also be
`inserted into the NMR probe by syringe flow-injection.
`They are dissolved in very small volumes that are
`matched to the active volume of the microcoil NMR probe
`(1.5 lL), thus further enhancing concentration and sensi-
`tivity.
`
`2 Materials and methods
`2.1 Materials and standards
`For HPLC, solvents (LiChrosolv gradient grade) from
`Merck (Darmstadt, Germany) and deionized water from a
`Milli-Q water purification system (Millipore AS, Bedford,
`MA, USA) were used. (all-E) Lutein, (all-E) zeaxanthin,
`(all-E) canthaxanthin, (all-E) b-carotene, and (all-E) lyco-
`pene were donated by BASF (Ludwigshafen, Germany).
`(all-E) Astaxanthin and b-apo-89-carotenal were pur-
`chased from Sigma (Steinheim, Germany). For conveni-
`ence, the trivial names of carotenoids are used throughout
`the text instead of the complex IUPAC nomenclature [34].
`Antarctic krill were fished on an expedition carried out by
`the Alfred Wegener Institute for Polar and Marine Research
`in the Wedell Sea, deep frozen with liquid nitrogen, and
`stored at – 408C until analysis. The extraction of the caro-
`tenoids from krill was performed utilizing MSPD with a
`silica based octadecyl sorbent material (C18 end-capped)
`from IST (Hengoed Mid Glam, UK) and butylated hydroxy-
`toluene as antioxidant from Sigma. Before analysis, the
`extract was sterile filtered through a syringe filter from
`Carl Roth (Karlsruhe, Germany). Iodine-catalyzed iso-
`merization of authentic (all-E) astaxanthin was performed
`similarly to Zechmeister [35]: An iodine solution was
`added to an (all-E) astaxanthin solution and the mixture
`was exposed to UV-light for 20 min. Lipase type VII from
`Candida rugosa and cholesterol esterase from Pseudo-
`monas flourescens for ester hydrolysis as well as bile salts
`were purchased from Sigma. Disodium hydrogen phos-
`phate, potassium dihydrogen phosphate, calcium chlor-
`ide, and sodium chloride were obtained from Merck. NMR
`measurements were carried out in acetone-d6 (99.8%)
`from Deutero GmbH (Kastellaun, Germany).
`
`2.2 Sample preparation
`2.2.1 Extraction
`The krill (0.5 g) was ground with 1.5 g of C18 (end-capped)
`MSPD material and butylated hydroxytoluene into a dry
`homogeneous powder. This takes between 5 and 10 min,
`depending on the amount of water in the sample. The
`force has to be adjusted to the sample as well: too much
`will destroy the silica particles and produce a high back
`pressure, too little will not be sufficient to generate a dry
`homogeneous powder. For quantitation, the sample was
`spiked with 10 lL of a b-apo-89-carotinal solution
`(13.03 mg/mL). The mixture was loaded into an empty
`
`SPE column and pressed between two frits. The column
`was washed with 10 mL of deionized water and the caro-
`tenoids were extracted with 4 mL of tert-butylmethyl ether
`(TBME). After evaporation of TBME under a nitrogen
`stream, the extract was stored at – 308C until it was ana-
`lyzed. Before analysis, the extract was redissolved in
`300 lL of TBME, sterile filtered, and the filter was washed
`three times with 500 lL of TBME. The solvent was evapo-
`rated again under nitrogen stream, and the extract was
`redissolved in 100 lL of ethanol, guaranteeing a high con-
`centration of the carotenoids.
`
`2.2.2 Enzyme-catalyzed ester hydrolysis
`The enzymatic assay was similar to the protocol devel-
`oped by Breithaupt [36]. In brief, 10 mL of phosphate buf-
`fer (0.1 M, pH 7.4), 30 mg of bile salts, and 250 lL of a cal-
`cium chloride/sodium chloride (75 mM/3 M) solution were
`preincubated with the dried MSPD extract for 30 min at
`378C. Then, 100 lL of a suspension of lipase (50 mg/mL)
`in a 5 mM calcium chloride solution was added and the
`mixture was incubated at 378C for 2 h. The carotenoids
`were extracted with chloroform, which was immediately
`evaporated under nitrogen stream. The residue was redis-
`solved in TBME, washed with water, dried with sodium
`sulfate, filtered, and the TBME was evaporated under
`nitrogen stream. Finally, the hydrolyzed extract was redis-
`solved in 100 lL of ethanol and subjected to HPLC anal-
`ysis.
`
`2.2.3 Transesterification of the carotenol fatty acid
`esters
`Two hundred microliters of 15% acetyl chloride in metha-
`nol was added to the extract and heated to 1108C for
`30 min. Then 100 lL of water was added and the esters
`extracted with 400 lL of n-hexane. The extract was dried
`with sodium sulfate and concentrated for GC MS analysis.
`
`2.3 Chromatography
`Analyses were carried out on an HP1100 system (Agilent
`Technologies, Waldbronn, Germany) using a UV detector
`(DAD) monitoring at 455 nm. The separations were per-
`formed on a 25064.6 mm ProntoSil C30 stainless steel
`column (Bischoff, Leonberg, Germany). The average
`pore diameter was 120 Š and the particle size was 3 lm.
`A C18 cartridge (Spark Holland BV, AJ Emmen, the Neth-
`erlands) was used as guard column. The separation of the
`carotenoids from the MSPD extract was achieved using a
`mobile-phase gradient elution program at a flow rate of
`1 mL/min. Two mixtures of methanol,
`tert-butylmethyl
`ether, and water were used as eluents (A – 83:15:2; v/v/v;
`B – 8:90:2; v/v/v). The elution started isocratically at
`100% A for 20 min, followed by a linear gradient to 40% A
`within 170 min. To verify the repeatability of the results,
`50 lL of the MSPD extracts were injected five times.
`
`J. Sep. Sci. 2005, 28, 1685 – 1693
`
`www.jss-journal.de
`
`i 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`RIMFROST EXHIBIT 1039 page 0003
`
`

`

`1688
`
`Grynbaum, Hentschel, Putzbach, Rehbein, Krucker, Nicholson, Albert
`
`2.4 HPLC MS
`MS was performed on a Bruker Esquire 3000plus LC-
`MS(n) system (Bruker Daltonics, Bremen, Germany)
`equipped with an APCI interface and an ion trap (IT). The
`HPLC APCI/MS coupling was accomplished using an
`HP1100 system (Agilent Technologies) and 50 lL of sam-
`ple was injected into the system. Mass spectra were
`recorded in the mass range of 50–1400 m/z. The detec-
`tion was performed using APCI in the positive ionization
`mode. The voltage of the corona needle was set to 3.5 kV.
`Nitrogen was used as drying gas as well as carrier gas at
`a flow rate of 4 L/min with a nebulizer pressure of 65 psi.
`The ionization chamber temperature was set to 3008C
`and the dry gas temperature was held at 2508C. The com-
`pound stability was set to 75% and the trap drive level to
`70%. The chromatographic conditions were the same as
`in the analytical separation described in Section 2.3.
`
`2.5 GC MS
`Analyses were performed on an HP 6890 system, coupled
`with an HP MD 5973 quadrupole mass spectrometer (Agi-
`lent Technologies). A fused-silica capillary Nordion SE54
`(25 m60.32 mm, 0.25 lm film thickness) was used for
`the separation. The carrier gas was helium with a constant
`flow rate of 1.3 mL/min. The injector temperature was set
`to 2808C and the column temperature program was as fol-
`lows: the initial temperature of 1008C was held constant
`for 2 min and then increased by 4 K/min to the final tem-
`perature of 2708C.
`
`2.6 NMR
`All NMR experiments were recorded using a Bruker AMX
`600 spectrometer
`(Bruker, Rheinstetten, Germany).
`Astaxanthin and two isomers were collected from two
`saponified MSPD extracts from the analytical HPLC sep-
`aration, as well as selected ester peaks from three MSPD
`extracts. The solvent was evaporated under nitrogen
`stream and the fractions were lyophilized and stored at
`– 308C until NMR analysis. For this, the fractions were
`redissolved in 15 lL of acetone-d6. 1H NMR spectra of the
`carotenoids were recorded utilizing syringe flow injection
`to a 1H selective microcoil NMR probe (Protasis/MRM,
`Savoy, IL, USA) with an active volume of 1.5 lL.
`
`1H NMR spectra were recorded with the pulse program zg
`without solvent suppression and a pulse program to sup-
`press the residual solvent signals (zgcpprsp) from deuter-
`ated acetone-d6 and water, using rectangular shaped
`pulses for low-power presaturation (rectangular pulses,
`length 100 ms). The temperature was set to 298 K. Tran-
`sients (4K) were recorded with a spectral width of 6024 Hz
`and 16K time domain points. The relaxation delay was set
`to 1 s. 1H,1H-COSY spectra were recorded for an unequi-
`vocal peak assignment with the pulse program cosy. Two
`
`hundred and fifty six transients with 2K complex data
`points and a spectral width of 6024 Hz were accumulated
`in the F2 dimension and 128 complex data points were
`accumulated in the F1 dimension.
`
`For all spectra, before Fourier transformation, a squared
`sine bell function was applied to the FID. Baseline correc-
`tion and phasing were performed manually. The chemical
`to acetone-d6,
`shift axis was referenced with respect
`d = 2.04 ppm, for all 1H NMR spectra. All NMR data were
`processed with XWIN-NMR version 3.5 (Bruker).
`
`3 Results and discussion
`Extraction of the carotenoids from krill was performed by
`MSPD. The deep frozen krill were ground with C18 sorbent
`material
`into a dry homogeneous powder, which was
`loaded into an SPE cartridge and pressed between two
`frits to a compact column bed. Polar compounds were
`eluted with water and the carotenoids were eluted with
`TBME. After sterile filtration, the solvent was evaporated
`and the dry residue was redissolved in 100 lL ethanol.
`The carotenoids were separated by HPLC on a C30 col-
`umn (25064.6 mm, 3 lm, 120 Š), resulting in the chro-
`matogram shown in Fig. 2. The chromatogram is divided
`into three parts. First, the more polar carotenoids elute. By
`UV spectroscopy (DAD) peak 2 is interpreted as (all-E)
`astaxanthin and peaks 1, 3, and 4 are interpreted as cis
`isomers. The absorption spectra of the two astaxanthin
`cis isomers (9-cis and 13-cis) are fairly similar to the one
`of the (all-E) isomer with only small differences [37–39]. In
`comparison to the (all-E) astaxanthin with a kmax of
`476 nm, the two isomers (peaks 1 and 4 in Fig. 2) show a
`small hypsochromic effect of 7 and 4 nm. The intensity of
`
`Figure 2. HPLC chromatogram (DAD, 455 nm) of the krill
`MSPD extract. Peaks 1–4 represent free astaxanthin, peak 5
`represents b-apo-89-carotinal (spiked), peaks 6–14 represent
`astaxanthin fatty acid monoesters, and peaks 15–31 repre-
`sent astaxanthin diesters. For precise peak assignment, see
`Table 2.
`
`J. Sep. Sci. 2005, 28, 1685 – 1693
`
`www.jss-journal.de
`
`i 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`RIMFROST EXHIBIT 1039 page 0004
`
`

`

`Astaxanthin and Astaxanthin Fatty Acid Esters in Krill
`
`1689
`
`Figure 3. HPLC chromatogram (DAD, 455 nm) of a standard
`solution containing astaxanthin (1), lutein (2), zeaxanthin (3),
`canthaxanthin (4), b-carotene (5), and lycopene (6).
`
`Figure 4. HPLC chromatogram (DAD, 455 nm) of krill MSPD
`extract that was subjected to enzymatic saponification using
`lipase from C. rugosa type VII.
`
`the cis band is greater since the cis double bond is nearer
`to the center of the molecule. This indicates that peak 1
`must be the (13-cis) isomer and peak 4 the (9-cis) isomer;
`this empirical assignment will be verified by NMR spectro-
`scopy. Peak 3 might be (15-cis) astaxanthin, as the UV
`spectrum was similar to the one of (all-E) astaxanthin
`(Dk = 1 nm), but this was not verified because the peak
`was too small for NMR analysis. Peak 5 is b-apo-89-carot-
`inal, which was used as internal standard for quantitation.
`It has passed through the same workup steps and does
`not interfere with the carotenoids extracted from krill. This
`means it is well suited for this purpose [40]. The fractions
`which elute later could be astaxanthin fatty acid esters; in
`this case, peaks 6–14 could be monoesters and peaks
`15–32 could be diesters.
`
`To identify carotenoids potentially contained in the krill by
`cochromatography, a carotenoid standard mixture was
`prepared and also subjected to HPLC under identical con-
`ditions (Fig. 3). All carotenoids are baseline separated
`and they elute according to their polarity in the following
`order: (all-E) astaxanthin (1; 11.6 min), (all-E) lutein (2;
`14.5 min),
`(all-E) zeaxanthin (3; 18.3 min),
`(all-E)
`canthaxanthin (4; 21.9 min),
`(all-E) b-carotene (5;
`68.5 min), and (all-E) lycopene (6; 154.5 min). (all-E)
`Astaxanthin was found to be the only one present in the
`extract. The other peaks were assumed to be carotenol
`fatty acid esters, the presence of which can be revealed
`by saponification.
`In addition, carotenoids that occur
`mainly in esterified form rather than in the free form can be
`revealed by that method. Under aerobic conditions, an
`alkali hydrolysis cannot be applied to esterified a-ketols
`like astaxanthin, because they undergo oxidation to form
`the 2,3-didehydro-3-hydroxy-4-keto end group in the pre-
`sence of base and oxygen [41]. For anaerobic saponifica-
`tion, a modified Schlenk tube with appendix can be used
`
`[42], but the procedure is rather involved. Enzymatic clea-
`vage of alkali-unstable carotenol esters has been applied
`successfully [36, 43]. The HPLC chromatograms of the
`saponified product and the nonsaponified samples are
`compared. If the eluted compound is more polar, but has
`an unchanged UV/VIS spectrum, that indicates that an
`ester was present and has been hydrolyzed.
`
`We successfully applied Lipase type VII from C. rugosa
`(Fig. 4) and cholesterol esterase from P. flourescens
`(data not shown) for the hydrolysis of the astaxanthin
`monofatty acid esters and difatty acid esters. No new
`peaks from carotenoids that might naturally occur only in
`their esterified form appear in the chromatogram of the
`hydrolyzed extract. The assumed astaxanthin monofatty
`acid esters and difatty acid esters (peaks 6–14 and peaks
`15–32, respectively) are almost completely hydrolyzed,
`while astaxanthin (peak 2) and the isomers (peaks 1, 3,
`and 4) dominate the chromatogram. Peak 2 shows the
`same UV/VIS spectrum as the esters with the absorption
`maximum at 476 nm. Due to the different solvent compo-
`sition, the absorption maximum of the esters is slightly
`shifted to 478 nm. Peaks 1–4 are magnified by a factor of
`9. This indicates that peaks 1, 3, and 4 are isomers of
`astaxanthin. The analytical proof was provided by iodine-
`catalyzed isomerization of the authentical standard. After
`isomerization under UV light, the mixture of (all-E) astax-
`anthin and the cis isomers was subjected to HPLC. The
`chromatogram showed the same isomer pattern as the
`krill extract (data not shown). Further MS and NMR
`experiments will show that these compounds are isomers
`and not products of decomposition.
`
`Next, GC MS experiments were performed to identify the
`fatty acids that are esterified to astaxanthin. For this, the
`MSPD extract was transesterified with acetyl chloride to
`the corresponding ethyl ester and the fatty acid esters
`
`J. Sep. Sci. 2005, 28, 1685 – 1693
`
`www.jss-journal.de
`
`i 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`RIMFROST EXHIBIT 1039 page 0005
`
`

`

`1690
`
`Grynbaum, Hentschel, Putzbach, Rehbein, Krucker, Nicholson, Albert
`
`Table 2. Peak assignment (astaxanthin and astaxanthin
`monoesters and diesters) for the peaks shown in Fig. 2.
`Peaks 1–4 are astaxanthin and its cis isomers, peak 5 is b-
`apo-89-carotinal (internal standard), peaks 6–14 are asta-
`xanthin fatty acid monoesters, and peaks 15–32 are asta-
`xanthin fatty acid diesters
`
`Peak
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`27
`28
`
`29
`30
`31
`32
`
`Assignment
`(13-cis) Astaxanthin
`(all-E) Astaxanthin
`Unidentified astaxanthin isomer
`(9-cis) Astaxanthin
`b-Apo-89-carotinal
`C12:0
`C14:0
`C16:1
`C14:0
`C18:1
`C16:0; C18:0; C18:1
`C14:0
`C16:0
`C16:0
`C12:0/C20:5
`C12:0/C12:0
`C12:0/C20:5
`C12:0/C22:6
`C12:0/C14:0
`C14:0/C22:6; C16:1/C22:6; C18:1/C20:5
`C18:1/C22:6
`C14:0/C20:5
`C12:0/C16:0; C14:0/C14:0; C12:0/C16:1
`C12:0/C16:0; C14:0/C14:0; C12:0/C16:1
`C12:0/C20:1; C14:0/C18:1; C16:1/C16:1
`C12:0/C20:1; C14:0/C18:1; C16:1/C16:1;
`C18:1/C18:1
`C12:0/C12:0
`C12:0/C18:1; C12:0/C18:0; C14:0/C16:0;
`C14:0/C16:1
`C16:0/C18:1; C16:1/C18:0; C16:1/C18:1
`C12:0/C18:0; C14:0/C16:0
`C16:1/C12:0
`C16:0/C16:0
`
`selected mass unit. Figure 6 exemplifies this with the
`selected ion chromatogram of astaxanthin myristic acid
`ester ([M + H]+ = 808 m/z). Three peaks are visible and
`their exact counterparts can be found in the UV chromato-
`gram. The largest peak at 55.9 min refers to the (all-E)
`astaxanthin fatty acid ester, while the ones at 50.0 and
`
`Figure 5. Gas chromatogram of the transesterified extract.
`For peak assignment, see Table 1.
`
`Table 1. Fatty acid composition of the krill MSPD extract
`
`Peak
`
`Fatty acid
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`
`C12:0
`C14:0
`C16:1
`C16:1
`C16:0
`C20:0
`C18:1
`C18:1
`C18:1
`C18:0
`C20:5
`C20:1
`C20:1
`C22:6
`
`Retention
`time, min
`15.23
`20.62
`25.05
`25.28
`25.60
`29.41
`29.55
`29.67
`29.88
`30.12
`33.04
`33.77
`33.91
`36.74
`
`Percen-
`tage,%
`0.1
`14.0
`6.6
`1.0
`25.4
`4.3
`17.0
`10.7
`0.7
`1.1
`12.0
`0.6
`0.5
`5.9
`
`were separated by GC on an SE-54 (5% phenyl/1% vinyl-
`methylpolysiloxane) column and identified by MS (Fig. 5).
`The most commonly occurring fatty acids are C14:0
`(14.0%), C16:0 (24.4%), C16:1 (7.6%), C18:1 (28.4%),
`C20:0 (4.3%), C20:5 (12.0%), and C22:6 (5.9%)
`(Table 1).
`
`With this information, the following step was to perform
`HPLC APCI/MS, which has been used in many previous
`experiments on carotenoids [25, 26] and carotenol fatty
`acid esters [27, 28], and has produced abundant proto-
`nated molecules [M + H]+ in the mass spectrum. Asta-
`xanthin, the geometrical isomers, and its monofatty acid
`esters and difatty acid esters were identified by SIM. For
`this, the mass spectrometer was set to scan over one
`
`J. Sep. Sci. 2005, 28, 1685 – 1693
`
`www.jss-journal.de
`
`i 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`RIMFROST EXHIBIT 1039 page 0006
`
`

`

`Astaxanthin and Astaxanthin Fatty Acid Esters in Krill
`
`1691
`
`64.1 min refer to the esters of geometrical isomers. Table
`2 depicts the assignment for most of the peaks from Fig. 2
`using this method. Several esters coeluted, such as the
`C16:0, C18:0, and C18:1 in peak 11, and some diesters
`were not distinguishable due to an identical mass, like
`C12:0/C16:0 and C14:0/C14:0 (peak 23). In contrast to a
`previous study [17], highly unsaturated fatty acids (C20:5,
`C22:6) were identified in astaxanthin esters, though only
`in diesters.
`
`Krill carapace and meat were also examined separately.
`The meat was found to be absolutely white, while the car-
`apace showed a strong magenta pigmentation, which
`indicates that the carotenoids are mainly stored in or
`directly under the carapace. HPLC APCI/MS analysis
`showed similar results, as no new peaks appeared and
`only the carotenoid content differed drastically, with
`clearly more carotenoid in the carapace. Separating the
`meat from the carapace was quite difficult, so this might
`also be due to insufficient partition.
`
`Structure elucidation of the geometrical isomers can be
`performed very efficiently by 1H NMR spectroscopy at
`high magnetic fields [38, 39, 44]. For this, peaks 1, 2, and
`4 were collected from two saponified extracts (1 g krill)
`and peaks 9, 10, 13, 28, 29, and 32 were collected from
`four extracts (2 g krill). Peak 3 could not be detected by
`NMR because it was too small.
`
`The position of the cis double bonds can be derived very
`reliably from the observed chemical shift changes
`between cis and trans isomers. Thus, stereomutation of
`(9-cis) or (13-cis) is revealed by distinct strong downfield
`shifts of about 0.5 ppm for H(8) and H(12). In contrast,
`protons attached to the convex side of the stereomutated
`bond are shifted upfield.
`
`The 1H,1H COSY is a very useful technique for the assign-
`ment of the carotenoids’ NMR signals, particularly in the
`crowded olefinic range. The chemical shifts and assign-
`ments for the (all-E) astaxanthin are compiled in Table 3.
`The signal H(7) is split by a vicinal trans coupling of
`
`Figure 6. Selected ion chromatogram for
`astaxanthin myristic acid ester
`([M + H]+ = 808 m/z).
`
`Table 3. Chemical shifts (d in ppm (acetone-d6)) and assign-
`ment of the 600 MHz 1H NMR signals of (all-E) astaxanthin,
`(13-cis) astaxanthin, and (9-cis) astaxanthin
`
`Protons
`
`H (7)
`H (79)
`H (8)
`H (89)
`H (10)
`H (109)
`H (11)
`H (119)
`H (12)
`H (129)
`H (14)
`H (149)
`H (15)
`H (159)
`
`all-E d,
`ppm
`6.35
`6.35
`6.50
`6.50
`6.38
`6.38
`6.77
`6.77
`6.52
`6.52
`6.40
`6.40
`6.79
`6.79
`
`13-cis Dd,
`ppm
`–
`–
`–
`–
`–
`0.01
`0.01
`– 0.01
`0.53
`0.01
`– 0.17
`0.06
`0.18
`– 0.04
`
`9-cis Dd,
`ppm
`–
`–
`0.58
`–
`– 0.04
`–
`0.17
`– 0.01
`– 0.05
`– 0.02
`0.01
`–
`–
`–
`
`16.2 Hz to H(8). The 3-spin subspectrum of H(10), H(11),
`and H(12) can be identified as doublet (H(10) 6.38 ppm),
`doublet of doublets (H(11) 6.77 ppm), and doublet (H(12)
`6.50 ppm) and their vicinal couplings J10,11 = 11.0 Hz and
`J11,12 = 14.0 Hz. Figure 7 depicts the olefinic region of the
`1H NMR spectra of (all-E) astaxanthin from krill, and of
`authentic (all-E) astaxanthin in acetone-d6. The signal at
`6.87 ppm in Fig. 7a derives from an impurity that was not
`detectable by UV at 455 nm and could not be separated
`from (all-E) astaxanthin.
`
`After the elucidation of (all-E) astaxanthin, two cis isomers
`(peaks 1 and 3 in Fig. 2) were investigated by NMR spec-
`troscopy. The Dd values, the shifts of the different cis and
`trans isomers’ proton signals, are listed in Table 3. The
`isomerization shift of the Proton 129 in the (13-cis) isomer
`has a downfield shift of Dd = 0.53 ppm and the H(149) has
`
`J. Sep. Sci. 2005, 28, 1685 – 1693
`
`www.jss-journal.de
`
`i 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
`
`RIMFROST EXHIBIT 1039 page 0007
`
`

`

`1692
`
`Grynbaum, Hentschel, Putzbach, Rehbein, Krucker, Nicholson, Albert
`
`erature, but no explanation can be provided. Astaxanthin
`and the fatty acid esters were detected unambiguously by
`HPLC MS, GC, and NMR. Other carotenoids were not
`detected.
`
`Future experiments could include online coupling of capil-
`lary HPLC with NMR. This sophisticated technique avoids
`any degradation that might occur during fraction collection
`as the sample is exposed to light and air, though this
`degradation was not observed in our experiments. The
`low solvent consumption in capillary HPLC allows the use
`of deuterated solvents. Still, the separation has to be
`adjusted to acetone and water, because they will – unlike
`the solvents applied in this work (TBME, methanol, and
`water) – only show two disturbing solvent signals in NMR
`spectra. However, the measurement time would be longer
`because of lower carotenoid concentration.
`
`Acknowledgments
`The authors would like to thank Dr. Martin Graeve from
`the Alfred-Wegener-Institute for Polar and Marine Re-
`search, Bremerhaven (Germany) for providing the krill
`sample and BASF AG, Ludwigshafen (Germany) for gra-
`tuitously providing the reference compounds (all-E) lutein,
`(all-E) zeaxanthin, (all-E) canthaxanthin, (all-E) b-caro-
`tene, and (all-E) lycopene. This work was supported by
`the Deutsche Forschungsgemeinschaft (AL 298/10-2).
`
`4 References
`[1] Braconnot, H., Ann. Chim