Fatty acid selectivity of microbial lipase and lipolytic enzymes from salmonid fish intestines toward astaxanthin diesters

Article abstract of DOI:10.1007/s11746-004-0905-8

The objective of the work described in this paper was to study a possible FA selectivity of digestive lipolytic enzymes isolated from salmon and trout intestines toward astaxanthin diesters of various FA composition and compare it with the FA selectivity of microbial lipase. Astaxanthin diesters of varying FA composition were prepared in excellent yields (>90%) by chemical esterification using a carbodiimide coupling agent. The astaxanthin diesters were screened in a hydrolysis reaction by various commercially available lipases. The highest conversion rates were observed with the Candida rugosa lipase, which discriminated against n-3 PUFA. The rate of hydrolysis was determined by HPLC. Digestive lipolytic enzymes isolated from salmon and rainbow trout intestines displayed reversed FA selectivity. Thus, astaxanthin diesters highly enriched with n-3 PUFA including EPA and DHA were observed to be hydrolyzed at a considerably higher rate than the more saturated esters. Similar trends in FA selectivity were observed in the hydrolysis of fish oil TAG by the digestive lipolytic enzyme mixtures.

Full text of DOI:10.1007/s11746-004-0905-8

Fatty Acid Selectivity of Microbial Lipase and Lipolytic
Enzymes from Salmonid Fish Intestines Toward
Astaxanthin Diesters
Arnar Halldorsson and Gudmundur G. Haraldsson*
Science Institute, University of Iceland, Dunhaga 3, 107 Reykjavik, Iceland
ABSTRACT: The objective of the work described in this paper deposited in the flesh of salmonids after hydrolysis of these
was to study a possible FA selectivity of digestive lipolytic en-
zymes isolated from salmon and trout intestines toward asta-
xanthin diesters of various FA composition and compare it with
the FA selectivity of microbial lipase. Astaxanthin diesters of
varying FA composition were prepared in excellent yields
(>90%) by chemical esterification using a carbodiimide cou-
pling agent. The astaxanthin diesters were screened in a hydrol-
ysis reaction by various commercially available lipases. The
highest conversion rates were observed with the Candida ru-
gosa lipase, which discriminated against n-3 PUFA. The rate of
hydrolysis was determined by HPLC. Digestive lipolytic en-
esters (5,9).
Astaxanthin is by far the most expensive feed ingredient on
a weight basis in salmonid fish farming (10), and normally
only 10–15% of the ingested astaxanthin can be recovered
from the muscle (5,6). There are several reasons for this very
low biological retention of commercial astaxanthin, including
various rather poorly understood processes related to diges-
tion and absorption (11). In addition, astaxanthin is very un-
stable with regard to exposure to air and elevated temperatures
as well as light. The pigment is therefore to a great extent de-
zymes isolated from salmon and rainbow trout intestines dis- graded during feed processing and storage. Clearly, even a
played reversed FA selectivity. Thus, astaxanthin diesters highly
enriched with n-3 PUFA including EPA and DHA were ob-
served to be hydrolyzed at a considerably higher rate than the
more saturated esters. Similar trends in FA selectivity were ob-
served in the hydrolysis of fish oil TAG by the digestive lipolytic
enzyme mixtures.
small improvement in bioavailability could result in substantial
savings for the salmonid aquaculture industry.
Astaxanthin as the diester is more stable than free asta-
xanthin (12). In the literature, dipalmitate is the predominant
diester studied, and it is reported to give less pigmentation
than the diol in both Atlantic salmon (Salmo salar) and rain-
bow trout (Oncorhyncus mykiss) (5,9,13). This is explained
by a low degree of hydrolysis of the diester in the fish in-
testines. However, in nature astaxanthin is often present as di-
ester. Crustaceans are the major source of carotenoids for
salmonids (9).
Paper no. J10691 in JAOCS 81, 347–353 (April 2004).
KEY WORDS: Astaxanthin, astaxanthin diesters, DHA, EPA,
lipase, lipolytic enzymes, n-3 PUFA, rainbow trout, salmon.
The main objectives of the study described in the current
paper were to investigate (i) whether astaxanthin diesters
were selectively hydrolyzed by a digestive lipolytic enzyme
mixture isolated from salmonid intestines and (ii) whether
such a possible selectivity could be reflected by astaxanthin
retention in salmon muscle. The answer to the first question
is provided by the results described in this report, whereas
the second question has been addressed in a related patent
publication (14).
Astaxanthins (3,3′-dihydroxy-β,β′-carotene-4,4′-diones; Fig. 1)
are carotenoid pigments of the xanthophyll type consisting of
eight isoprene units. Astaxanthins are abundant in nature and
can be found in fish, crustaceans, and birds, as well as in cer-
tain algae, plants, fungi, and bacteria (1–3). They are synthe-
sized de novo by algae and plants as well as bacteria and
fungi. In algae, phytoplankton, and plants, carotenoids, in-
cluding astaxanthins, play key roles in photosynthesis and
photoprotection. Astaxanthin is responsible for much of the
red coloring in the animal kingdom, e.g., in various crus-
taceans such as lobster, shrimp, and crab and in the spawning
array and flesh of many fish including salmon and trout. As-
taxanthin is of great significance for both the health and re-
productive capability of these fish.
Astaxanthin is the major carotenoid of wild salmon and the
most commonly used carotenoid for pigmentation in salmonid
fish farming (4–6). The most important commercial source of
astaxanthin is a synthetic material comprising a rather compli-
cated mixture of astaxanthin geometrical and optical isomers
(7). It is supplied in an unesterified form, which is the de-
posited form of the pigment in the fish muscle (5). Natural
sources of astaxanthin, such as krill or algae, yield astaxanthin
predominantly in the esterified form (8). Free astaxanthin is
EXPERIMENTAL PROCEDURES
Instrumentation. 1H and 13C NMR spectra were recorded on
a Bruker AC 250 NMR spectrometer in deuterated chloro-
form as a solvent. FA analyses were performed on methyl es-
ters employing a PerkinElmer 8140 gas chromatograph using
a 30-m capillary column, DB-225 30N 0.25 mm (J&W Sci-
entific, Folsom, CA) with hydrogen as a carrier gas according
to our previously described procedure (15). Analytical HPLC
was performed on a constaMetric 3200 solvent delivery sys-
tem with spectroMonitor 3200 variable wavelength detector
(470 nm), both from LDC Analytical using a Nucleosil
50-5 column (Art. 721210) from Macherey-Nagel (Easton,
PA). Electronic spectra were obtained on an Ultrospec
4000 UV/vis spectrophotometer from Pharmacia Biotech.
* To whom correspondence should be addressed. E-mail: gghar@raunvis.hi.is
Copyright © 2004 by AOCS Press
347
JAOCS, Vol. 81, no. 4 (2004)

348
A. HALLDORSSON AND G.G. HARALDSSON
FIG. 1. Structures of the three stereoisomers of all-E astaxanthin: (3S,3’S), (3R,3’R), and meso.
Centrifugation was performed on a Sorval RC5C centrifuge from Sigma-Aldrich Chemical Company (St. Louis, MO).
and ultracentifugation on a Beckman L5-50E centrifuge. Ag- They were all used without further purification. Solvents
itation was performed on a VXR basic IKA-Vibrax unless (n-hexane, dichloromethane, and acetone) were obtained from
otherwise stated.
Acros Organics (Geel, Belgium) and were of analytical grade
Materials and solvents. The microbial lipases were supplied and used without further purification. Free astaxanthin, with
and employed as powder by Amano Enzyme Europe Ltd. (Mil- an indicated extinction coefficient of 2070 in chloroform (1%
ton Keynes, England). They include Pseudomonas cepacia (Li- solution, 1-cm path) at 492 nm, was obtained as a crystalline
pase PS), P. fluorescens (Lipase AK), Geotrichum candidum material from Alexis Biochemicals (San Diego, CA). The
(Lipase GC), Penicillium roqueforti (Lipase R), Aspergillus polyoxyethyleneglycol (PEG) emulsifier was supplied as Cre-
niger (Lipase A), Candida rugosa (Lipase AE), Humicula mophor EL by BASF AG (Ludwigshafen, Germany). Herring
lanuginosa (Lipase CE), Rhizopus delemar (Lipase D), R. oil was from Pronova Biocare AS (Sandefjord, Norway). Cod
oryzae (Lipase F), P. camembertii (Lipase G), C. lipolytica (Li- liver oil was obtained from Lysi hf (Reykjavik, Iceland).
pase L), Rhizomucor javanicus (Amano M), and R. niveus li- Farmed Atlantic salmon (S. salar) were obtained live (3–5
pase (Lipase N). Pig liver esterase was purchased from Sigma kg) from Islandslax hf (Grindavik, Iceland), and farmed
Chemical Company (St. Louis, MO). EPA (98%), DHA rainbow trout (O. mykiss) were obtained live (1–1.5 kg)
(>95%), astaxanthin dicaprate and dielaidate, and concentrates from Stofnfiskur hf (Hafnarfjordur, Iceland).
of various EPA/DHA composition, including 85% (50% EPA,
Preparation of astaxanthin diesters. The free astaxanthin
35% DHA), 55% (33% EPA, 22% DHA), and 30% (18% EPA, starting material displayed the following NMR spectrum: 1H
12% DHA), were obtained as free acids from Norsk Hydro Re- NMR δ 6.70–6.61 (m, 4H, =CH), 6.48–6.18 (m, 10H, =CH),
search Centre (Porsgrunn, Norway). Reagent grade tris[hy- 4.33 (dd, J = 5.6 Hz, J = 13.8 Hz, 2H, CHOH), 3.63 (br s, 2H,
droxymethyl]aminomethane (Trisma base, 99.9%) and OH), 2.16 (dd, 5.6 Hz, J = 12.7 Hz, 2H, CH2CHOH), 2.00
palmitic acid (99%) were purchased from Sigma Chemicals. (s, 6H, CH3C=CH), 1.99 (s, 6H, CH3C=CH), 1.95 (s, 6H,
Silica gel (Silica gel 60), analytical TLC plates (DC Alufolien =C(CH3)CO), 1.81 (dd, J = 12.7 Hz, J = 13.8 Hz, 2H,
Kieselgel 60 F254), 4-dimethylaminopyridine (DMAP, >99%) CH2CHOH), 1.32 (s, 6H, C(CH3)2), and 1.21 (s, 6H,
and hydrochloric acid (37% w/w) were obtained from Merck C(CH3)2) ppm.
(Darmstadt, Germany), and 1-(3-dimethylaminopropyl)-3-
Astaxanthin dipalmitate. To a mixture of pure astaxanthin
ethylcarbodiimide hydrochloride (EDCI, >98%) was obtained (251 mg, 0.421 mmol) and palmitic acid (220 mg, 0.858
JAOCS, Vol. 81, no. 4 (2004)

ASTAXANTHIN DIESTERS
349
mmol) in dichloromethane (5 mL) was added DMAP (100
mg, 0.820 mmol) and EDCI (200 mg, 1.043 mmol), and the
resulting solution was stirred on a magnetic stirrer at room
temperature (r.t.) for 24 h. The reaction was monitored by
TLC using 95:5 (vol/vol) dichloromethane/acetone as an elu-
ent (diol: Rf 0.0–0.1, monoester: Rf 0.3–0.5, diester: Rf
0.8–0.9). The solvent was evaporated off in vacuo on a rotary
evaporator and the crude reaction mixture run through a short
silica gel column using dichloromethane as an eluent. Pure
palmitate diesters were obtained as a deep red-purple powder
Astaxanthin diesters of 55% EPA + DHA concentrate. An
identical procedure was used as described for the DHA
adduct above using astaxanthin (250 mg, 0.419 mmol) and
the 55% EPA + DHA concentrate as free acids (259 mg, ap-
proximately 0.863 mmol) in dichloromethane (5 mL), DMAP
(101 mg, 0.823 mmol), and EDCI (199 mg, 1.038 mmol). The
55% EPA + DHA diesters were obtained as a deep red-purple
syrupy material (457 mg, 94% yield).
Astaxanthin diesters of 30% EPA + DHA concentrate. An
identical procedure was used as described for the DHA
adduct above using astaxanthin (252 mg, 0.422 mmol) and
the 30% EPA + DHA concentrate as free acids (251 mg, ap-
proximately 0.865 mmol) in dichloromethane (5 mL), DMAP
(102 mg, 0.835 mmol), and EDCI (198 mg, 1.033 mmol). The
30% EPA + DHA diesters were obtained as a deep red-purple
syrupy material (453 mg, 94% yield).
Astaxanthin diester stock solution. A Tris buffer (0.25 M)
solution was prepared by dissolving Tris (4.54 g, 37.5 mmol)
in 150 mL distilled water. The pH of the buffer solution was
adjusted to 8.0 by adding 2 M hydrochloric acid. All the hy-
drolysis reactions were performed at r.t. so no special effort
to adjust the pH was necessary. The astaxanthin diester was
weighed accurately (approximately 40 mg, 33 mmol) along
with Chremophor EL emulsifier (2.5 g) into a 100-mL Erlen-
meyer flask. The solution was stirred vigorously for 20 min
(until homogeneous), and then 20 mL of the Tris buffer was
added and the mixture was stirred for an additional 10 min.
The solution was suction-filtered, and the red filtrate was
poured into a 25-mL volumetric flask and diluted to the mark
with buffer. The concentration of all the solutions was deter-
mined spectrophotometrically (492 nm) to be about 1 mg/mL.
All the solutions were stored at 4°C under nitrogen and pro-
tected from light.
1
(415 mg, 92% yield). H NMR δ 6.71–6.60 (m, 4H, =CH),
6.48–6.17 (m, 10H, =CH), 5.53 (dd, J = 6.2 Hz, J = 13.0 Hz,
2H, CHOCO), 2.54–2.33 (m, 4H, CH2COO), 2.18–1.94 (m,
4H, CH2CHOCO), 2.00 (s, 6H, CH3C=CH), 1.99 (s, 6H,
CH3C=CH), 1.90 (s, 6H, =C(CH3)CO), 1.76–1.64 (m, 4H,
CH2CH2COO), 1.38–1.20 (m, 48H, CH2), 1.35 (s, 6H,
C(CH3)2), 1.22 (s, 6H, C(CH3)2), and 0.88 (t, 6H, J = 6.6 Hz,
CH3CH2) ppm.
Astaxanthin dieicosapentaenoate. An identical procedure
was used as described for the palmitate adduct above using
astaxanthin (250 mg, 0.419 mmol) and EPA as free acid
(259 mg, 0.855 mmol) in dichloromethane (5 mL), DMAP
(102 mg, 0.835 mmol) and EDCI (197 mg, 1.028 mmol). Pure
EPA diesters were obtained as a deep red-purple syrupy ma-
1
terial (449 mg, 92% yield). H NMR δ 6.70–6.60 (m, 4H,
=CH in astax.), 6.47–6.16 (m, 10H, =CH in astax.), 5.53 (dd,
J = 6.2 Hz, J = 13.1 Hz, 2H, CHOCO), 5.43–5.28 (m, 20H,
=CH in EPA), 2.86–2.78 (m, 16H, =CCH2C=), 2.52–2.37 (m,
4H, CH2COO), 2.20–1.94 (m, 12H, CH2CH2C=, CH3CH2C=,
and CH2CHOCO), 1.99 (s, 6H, CH3C=CH), 1.98 (s, 6H,
CH3C=CH), 1.89 (s, 6H, =C(CH3)CO), 1.83–1.72 (m, 4H,
CH2CH2COO), 1.34 (s, 6H, C(CH3)2), 1.21 (s, 6H, C(CH3)2),
and 0.97 (t, J = 7.5 Hz, 6H, CH3CH2C=) ppm.
Astaxanthin didocosahexaenoate. An identical procedure
was used as described for the palmitate adduct above using as-
taxanthin (250 mg, 0.419 mmol) and DHA as free acid (282
mg, 0.858 mmol) in dichloromethane (5 mL), DMAP (103
mg, 0.840 mmol), and EDCI (200 mg, 1.043 mmol). Pure
DHA diesters were obtained as a deep red-purple syrupy ma-
terial (464 mg, 91% yield). 1H NMR δ 6.70–6.61 (m, 4H, =CH
in astax.), 6.47–6.17 (m, 10H, =CH in astax.), 5.53 (dd, J =
6.2 Hz, J = 13.0 Hz, 2H, CHOCO), 5.44–5.32 (m, 24H, =CH
in DHA), 2.87–2.79 (m, 20H, =CCH2C=), 2.52–2.43 (m, 8H,
CH2CH2COO), 2.18–1.95 (m, 8H, CH3CH2C= and
CH2CHOCO), 2.00 (s, 6H, CH3C=CH), 1.99 (s, 6H,
CH3C=CH), 1.90 (s, 6H, =C(CH3)CO), 1.35 (s, 6H, C(CH3)2),
1.22 (s, 6H, C(CH3)2), and 0.97 (t, J = 7.5 Hz, 6H,
CH3CH2C=) ppm.
Astaxanthin diesters of 85% EPA + DHA concentrate. An
identical procedure was used as described for the DHA
adduct above using astaxanthin (251 mg, 0.421 mmol) and
the 85% EPA + DHA concentrate as free acids (264 mg, ap-
proximately 0.860 mmol) in dichloromethane (5 mL), DMAP
(100 mg, 0.818 mmol), and EDCI (198 mg, 1.033 mmol). The
85% EPA + DHA diesters were obtained as a deep red-purple
syrupy material (455 mg, 92% yield).
Isolation of the lipolytic enzyme preparation. Fresh salmon
and rainbow trout that had been fed in the last 12 h were ob-
tained live from local fish farms. The fresh fish was cut open
at the laboratory and the crude enzyme mixture extracted as
follows: Each of the blind-ending tubes (pyloric caecae) orig-
inating from the pyloric area of the stomach were cut off, and
the intestinal juice was squeezed out by hand into a cooled
container. The viscous solution (approximately 50 mL) was
diluted to 100 mL with 0.25 M Tris buffer, pH 8.0, stirred for
30 min in an ice bath and centrifuged at 12,000 × g (rotor
GSA) for 20 min at 4°C. The aqueous layer was removed to
another centrifuge tube and the centrifugation repeated at
25,000 × g (rotor SS-34) for 30 min at 4°C. The aqueous layer
was transferred to an Erlenmeyer flask and quickly frozen by
liquid nitrogen. In some instances, before use the mixture was
thawed and purified by ultracentrifugation at 100,000 × g
(rotor SW27) for 45 min at 4°C. The resulting enzyme prepara-
tion was a clear yellow solution.
Hydrolysis of astaxanthin diesters. Into a 10-mL round-
bottomed flask was added 2 mL of the astaxanthin diester
stock solution and 3 mL of lipolytic enzyme mixture from fish
intestine. The flask was flushed with nitrogen before closing
and then wrapped in aluminum foil, and the reaction mixture
JAOCS, Vol. 81, no. 4 (2004)

350
A. HALLDORSSON AND G.G. HARALDSSON
was stirred on a magnetic stirrer at r.t. for 48 h. When fin-
The most important commercial source of astaxanthin is
ished, all the water was removed under reduced pressure Carophyll Pink from Hoffmann-La Roche (Basel, Switzer-
(0.01 Torr) at r.t. and the residue redissolved in land). This synthetic material contains approximately 75% of
dichloromethane. The solution was filtered through a cotton the natural all-E isomer, roughly as a 50:50 mixture of
wool plug in a Pasteur pipette and stored under nitrogen in a racemic optically inactive enantiomers and meso-astaxan-
closed
con- thins (Fig. 1). This means approximately 1:1:2 of the (R,R′)-
tainer protected from light in a freezer. Analytical TLC was and (S,S′)-enantiomers and the meso (R,S)-diastereomer, re-
used to monitor the progress of the reaction with 95:5 spectively. The remaining 25% consists of Z-geometrical iso-
(vol/vol) dichloromethane/acetone as an eluent. To determine mers, of which the 9Z, 13Z, and 15Z isomers are the most
the degree of hydrolysis, samples were subjected to analyti- prominent (7). In the current studies, a synthetic astaxanthin
cal HPLC using 70:30 (vol/vol) n-hexane/acetone as an elu- from Alexis Biochemicals (San Diego, CA) was used. It is
ent (flow rate: 0.15 mL/min; detector: 470 nm. Approximate virtually free of any Z-geometrical isomers and contains a
retention times: astaxanthin diesters: 9 min; astaxanthin similar ratio of the all-E isomers as the Carophyll Pink asta-
monoesters: 12 min; astaxanthin diols: 22 min.)
xanthin material.
Astaxanthin diester hydrolysis studies by pig liver esterase
Lipases are known to tolerate a wide range of primary and
and microbial lipase. Lipase or pig liver esterase (20 mg) was secondary alcohols as substrates (16), and astaxanthins are
placed in a screw-capped reaction vessel together with asta- among substrates reported to be tolerated by lipase (17). The
xanthin dielaidate stock solution (2 mL) under a nitrogen at- screening of numerous commercially available microbial lipases
mosphere. A glass rod was used to rub lumps of enzyme into and pig liver esterase was necessary for numerous reasons:
solution. The reaction vessel was wrapped in aluminum foil first, to find and use a well-defined esterase or microbial li-
to protect the mixture from light and then placed on a me- pase to serve as a model to set the methodology for the fish
chanical stirrer (IKA-Vibrax-VXR); the reaction was allowed intestinal lipolytic enzyme mixture, and second, to provide a
to proceed for 48 h at r.t. The workup was identical to that in comparison between such an esterase or lipase, once found
the hydrolysis reaction with the lipolytic salmonoid enzyme active, and the fish-based lipolytic enzymes in their asta-
mixture described above.
xanthin ester hydrolysis in terms of FA selectivity. Finally,
the results with such a lipase or esterase could possibly pro-
vide good evidence together with enzyme-free reactions, that
the results with the fish enzymes were not an artifact related
RESULTS AND DISCUSSION
The bioavailability of astaxanthin for salmon and trout in to the instability of astaxanthin diesters highly enriched with
aquaculture is a complicated issue that relates to both stabil- the labile EPA and DHA.
ity and various poorly understood digestion and absorption
Preparation of astaxanthin diesters. In the studies described
processes (11). In nature, astaxanthin is commonly present in this paper, astaxanthin from Alexis Biochemicals was used.
as diesters, and esterification is known to increase stability From that material astaxanthin diesters were prepared by chem-
(12). However, astaxanthin dipalmitates are largely inferior ical esterification with capric, palmitic, and elaidic acids, con-
to free astaxanthins in terms of pigment retention in salmon centrates of varying n-3 PUFA content, and virtually pure EPA
and trout muscle (5,9,13). This is believed to relate to a lower and DHA. The concentrates include 30% EPA + DHA (18%
degree of hydrolysis of such diesters in the fish intestines. It EPA, 12% DHA), 55% EPA + DHA (33% EPA, 22% DHA),
was anticipated that the FA composition of astaxanthin di- and 85% EPA + DHA (50% EPA, 35% DHA), the last one con-
esters might influence the degree of hydrolysis in the fish in- taining greater than 90% total n-3 PUFA content. The reaction
testines as well as their stability and bioavailability.
is schematically outlined in Scheme 1.
The main objective of the work described in this paper was to
The reactions were conducted in dichloromethane at r.t.
study a possible FA selectivity of digestive lipolytic enzymes iso- using EDCI as a coupling agent in the presence of DMAP.
lated from salmon and rainbow trout intestines toward asta- EDCI is a coupling agent similar to dicyclohexylcarbodiimide
xanthin diesters of various FA compositions. For this purpose, a but far superior when n-3 PUFA are involved (18). Its polarity
lipolytic enzyme mixture from salmonid intestines was required, and water solubility render it quite an advantage in terms of
and its activity had to be established in vitro on astaxanthin di- workup of the reaction and purification of the products.
esters. Astaxanthin diesters of various FA compositions, includ- DMAP is presumed to serve both as a base and a catalyst for
ing n-3 PUFA, also had to be prepared.
the acylation process. The same procedure was previously
SCHEME 1
JAOCS, Vol. 81, no. 4 (2004)

ASTAXANTHIN DIESTERS
351
used successfully to introduce pure EPA and DHA into the
mid-position of 1,3-DAG (18,19).
Astaxanthin diester hydrolysis by C. rugosa lipase. Asta-
xanthin diester hydrolysis by C. rugosa lipase was conducted on
diesters of capric, palmitic, and elaidic acids, and the 85% EPA
+ DHA concentrate. The reactions were followed by analytical
TLC as before. They were discontinued after 42 h and analyzed
by analytical HPLC. The results are presented in Table 1.
The diesters were obtained in virtually quantitative yields
and isolated in well over 90% yields after purification by silica
1
gel chromatography. High-field H NMR spectroscopy was
used to evaluate the high degree of purity of the products. Al-
though the astaxanthin comprised a mixture of stereoisomers,
the protons located α to the carbonyl group of the cyclo-
hexenone moiety, on the carbon possessing the hydroxyl group,
resonated as a regular distinct doublet of doublets at δ 4.43 ppm.
Upon acylation, a dramatic downfield shift of these protons took
place to δ 5.53 ppm. This, together with a complete disappear-
ance of the hydroxyl group proton signal, was good evidence
for the formation and high purity of the astaxanthin diacyl
adducts. The remaining regions of the spectra were also in ex-
cellent agreement with the postulated introduction and identity
of two acyl group equivalents into the astaxanthin molecule.
Hydrolysis reactions. The astaxanthin diester hydrolysis
reactions (Scheme 2) were conducted at r.t. in a Tris buffer at
pH 8 using a PEG emulsifier supplied as Cremophor EL by
BASF. HPLC possessing a variable wavelength detector fixed
at 470 nm was used to analyze the product mixture in terms
of unreacted diesters, monoesters, and free astaxanthins
(diols), from which the extent of conversion was determined.
The percentage conversion (%conv.) was defined and calcu-
lated according to the following equation:
The Candida lipase clearly displayed preference for the
saturated and less unsaturated fatty adducts as compared to
the n-3 PUFA concentrate, as had been anticipated. That li-
pase has previously been shown to discriminate against n-3
PUFA in fish oil, DHA in particular (20,21). By far the high-
est degree of hydrolysis (75%) was observed with the elai-
date esters. It is interesting to notice the low content of mo-
noesters in that case, much lower than for the other esters.
Preliminary astaxanthin diester hydrolysis by salmon in-
testine lipolytic enzymes. When hydrolysis studies similar to
those described above for the Candida lipase were conducted
on an enzyme mixture isolated from salmon intestines, a re-
versed behavior was clearly observed in terms of FA selectiv-
ity (Table 2). This time the highest activity was observed for
the n-3 PUFA concentrate: 41% conversion after 45 h. This
compares to 14% for the caprate esters and only 4 and 6% for
the palmitate and elaidate esters, respectively. It is notewor-
thy that a substantial amount of free astaxanthin was pro-
duced from the n-3 PUFA astaxanthin diesters.
Astaxanthin diester hydrolysis by salmon and rainbow trout
intestine lipolytic enzymes. For better insight into the FA selec-
tivity of the crude enzyme mixture from salmon intestines, asta-
xanthin diesters prepared from FA mixtures containing various
levels of EPA and DHA ranging from 30% EPA + DHA to pure
EPA and DHA were investigated. A fresh enzyme mixture from
freshly killed salmon was used. The palmitate diester was in-
cluded for comparison. The results are presented in Table 3. It is
noteworthy that the conversion levels were higher than those
obtained in the previous experiment (Table 2). This time, how-
ever, the diol levels were significantly lower than the monoester
%monoesters
%conv. = %diols +
[1]
2
The percentage was based on area percentage on HPLC and
represents molar percentage. Since the astaxanthin deposited
into fish muscle is free astaxanthin, we were particularly in-
terested in the diol content of these reactions, especially when
it came to the fish-derived enzymes.
Screening of enzymes. The hydrolytic activities of numerous
commercially available lipases (listed in the Experimental Pro-
cedures section) and pig liver esterase toward astaxanthin di-
esters of elaidic acid were determined. Analytical TLC on silica
gel was used to follow the progress of the hydrolysis reactions.
Less than half of the enzymes examined displayed some
activity toward the diesters, mostly generating monoesters.
Active enzymes included both Pseudomonas lipases tested,
and the G. candidum, C. rugosa, Rhizopus delemar, R.
oryzae, and P. roqueforti lipases. The highest conversion rate
TABLE 1
Astaxanthin Diester Hydrolysis by Candida rugosa Lipase (42 h)a
Conv.
(%)
Diesters
(%)
Monoesters
(%)
Diols
(%)
Substrate
10:0
16:0
18:1
K85
48
50
75
40
34
41
24
39
36
19
3
30
40
73
18
43
was observed for the C. rugosa lipase, which also was the aMolar percentages of the astaxanthin diesters, monoesters, and diols were
only lipase to produce diols in substantial amounts. There-
fore, that lipase became an obvious choice for further hydrol-
determined by HPLC using detection at 470 nm. Conv., conversion calcu-
lated as the sum of astaxanthin diols and half of the astaxanthin monoesters;
10:0, capric acid; 16:0, palmitic acid; 18:1 elaidic acid (trans-18:1); K85,
ysis studies.
85% EPA + DHA in approximately 3:2 ratio, respectively.
SCHEME 2
JAOCS, Vol. 81, no. 4 (2004)

352
A. HALLDORSSON AND G.G. HARALDSSON
TABLE 2
TABLE 4
Astaxanthin Diester Hydrolysis by Salmon Intestinal Lipolytic
Astaxanthin Diester Hydrolysis by Rainbow Trout Intestine Lipolytic
Enzymes (45 h)a
Enzymes (48 h)a
Conv.
(%)
Diesters
(%)
Monoesters
(%)
Diols
(%)
Conv.
(%)
Diesters
(%)
Monoesters
(%)
Diols
(%)
Substrate
Substrate
10:0
16:0
18:1
K85
14
4
6
78
94
90
43
16
5
9
6
1
1
16:0
K30
K55
K85
EPA
DHA
20
40
49
57
54
52
71
37
25
14
17
21
18
46
52
59
58
55
11
17
23
27
25
24
41
31
25
aSee Table 1 for abbreviations.
aSee Tables 1 and 3 for abbreviations.
levels, using K85 as substrate. The difference in performance be-
tween the two extracts presumably relates to an extra purifica- aluminum foil, and all experiments were conducted under a
tion step by ultracentrifugation as described above in “Isolation nitrogen atmosphere. There was no apparent deterioration in
of the lipolytic enzyme preparation,” resulting in higher purity color intensity of the astaxanthin solutions during the experi-
as compared to the previous experiment. This may indicate that ments and not much reason to believe that oxidation was tak-
more than a single hydrolytic enzyme was responsible for the ing place during the reactions to any harmful extent.
hydrolytic activity, possibly lipases, esterases, or an astaxanthin
Fish oil hydrolysis by salmon and rainbow trout intestine
esterase. This requires further studies, but the trends clearly show lipolytic enzymes. We were also interested in investigating
a correlation between enzymatic activity and n-3 PUFA content whether similar trends in FA selectivity of the salmon and trout
of the diester substrate, with the 85% EPA + DHA diester and lipolytic enzymes were displayed in the hydrolysis of fish oil
pure EPA diester offering the highest conversion.
TAG. When herring oil was incubated with the crude enzyme
Similar results were obtained for the rainbow trout enzyme mixture from salmon intestines under similar conditions as
mixture (Table 4). Significantly higher degrees of hydrolysis those described for the astaxanthin diesters, approximately 10%
were obtained for all astaxanthin diesters as compared to the conversion was obtained after 72 h. The hydrolytic enzymes
salmon enzymes. Again, the 85% EPA + DHA concentrate di- clearly preferred the long-chain n-3 PUFA as substrates over
esters displayed the highest activity and conversion (57%) as saturated and less unsaturated FA, unlike commercially
well as the highest amount of free astaxanthins (27%). From available lipases (22).
these results it is evident that astaxanthin diesters containing high
levels of EPA and DHA are preferred to diesters of saturated and oil by the crude enzyme mixture from rainbow trout intestines.
less unsaturated FA as substrates for the trout enzyme mixture.
Much higher hydrolysis (25%) was observed under identical
Similar results were obtained for the hydrolysis of cod liver
No attempts were made to look into the possible selectivity conditions. As before, the enzyme mixture displayed a strong
of the enzymes toward individual stereoisomers of astaxanthin preference for EPA and DHA and other n-3 PUFA as compared
diesters and their possible enantiopreference. This is under in- to the saturated and less unsaturated FA (23). A question remains
vestigation, but the results observed in this study uniformly indi- as to whether the hydrolytic activity involving both astaxanthin
cate that the enzymes prefer the n-3 PUFA-enriched astaxanthin diesters and the fish oil TAG is solely related to lipase or to both
esters as substrates irrespective of their enantiopreference and lipase and esterase. Whatever the answer to this question, simi-
stereoisomeric composition of the astaxanthins. The stability lar trends in FA selectivity were displayed in both cases. A simi-
of the free astaxanthins and their highly enriched EPA and lar FA selectivity favoring n-3 PUFA was reported by Lie and
DHA esters during the hydrolysis reactions was of concern, Lambertsen (24) in their studies on fish oil TAG hydrolysis using
although no attempts were made to evaluate their oxidative a crude hydrolytic enzyme mixture from cod intestines. Finally,
stability. The hydrolysis experiments were performed with a bile-salt-dependent lipase from cod intestines was observed
great care, with the reaction vessels protected from light by by Gellesvik (25) to display a similar preference for long-chain
n-3 PUFA up to 22 carbons in length.
TABLE 3
Feeding experiments on salmon. In vivo studies based on
the results described in this report were conducted by Norsk
Hydro in Norway on farmed salmon and revealed that asta-
xanthin deposition in fish muscle was considerably higher
when using astaxanthin diesters highly enriched with n-3
PUFA as compared to free astaxanthin (14). For the asta-
xanthin diesters prepared from 85% EPA + DHA and greater
than 90% total n-3 PUFA, 19.7% of injected astaxanthin was
deposited into fish muscle. This compares to 13.0% for fish
receiving free astaxanthin, corresponding to a 50% increase
in astaxanthin bioavailability. This appears to be related to the
preferential activity of the hydrolytic enzymes (esterase or li-
Astaxanthin Diester Hydrolysis by Salmon Intestinal Lipolytic
Enzymes (48 h)a
Conv.
(%)
Diesters
(%)
Monoesters
(%)
Diols
(%)
Substrate
16:0
K30
K55
K85
EPA
DHA
10
21
32
36
40
30
81
61
42
37
27
44
19
36
52
54
67
53
0
3
6
9
6
3
aK30, K55, and K85 are n-PUFA concentrates comprising 30, 55, and 85%
EPA + DHA, respectively, in approximately 3:2 ratio. See Table 1 for other
abbreviations.
JAOCS, Vol. 81, no. 4 (2004)

ASTAXANTHIN DIESTERS
353
Carotenoid Oil Extracted from Antarctic Krill, Aquaculture
66:255–264 (1987).
pase) in the fish intestines for such diesters. Such an in vivo
hydrolysis appears to increase astaxanthin bioavailability in
the fish as compared to the feeding of free astaxanthin.
13. Storebakken, T., P. Foss, K. Schiedt, E. Austreng, S. Liaaen-
Jensen, and U. Manz, Carotenoids in Diets for Salmonids. IV.
Pigmentation of Atlantic Salmon with Astaxanthin, Astaxanthin
Dipalmitate and Canthaxanthin, Ibid. 65:279–292 (1987).
14. Breivik, H., L.I. Sanna, and B.A. Aanesen, Pigment, WO
00/62625 (1999).
15. Haraldsson, G.G., and Ö. Almarsson, Studies on the Positional
Specificity of Lipase from Mucor miehei During Interesterifi-
cation Reactions of Cod Liver Oil with n-3 Polyunsaturated
Fatty Acid and Ethyl Ester Concentrates, Acta Chem. Scand.
45:723–730 (1991).
ACKNOWLEDGMENTS
Dr. Harald Breivik and Berit Aanesen at Norsk Hydro Research Cen-
tre in Porsgrunn, Norway, are acknowledged for collaboration, as is
Dr. Sigridur Jonsdottir at the Science Institute for running the NMR
analysis, and Norsk Hydro in Porsgrunn for providing virtually pure
EPA and DHA, some of the astaxanthin diesters, and financial support.
16. Haraldsson, G.G., The Application of Lipases in Organic Syn-
thesis, in The Chemistry of the Functional Groups, Supplement
B2: The Chemistry of Acid Derivatives, edited by S. Patai, John
Wiley & Sons, Chichester 1992, Vol. 2, pp.1395–1473.
17. Nagao, T., T. Fukami, Y. Horita, S. Komemushi, A. Sugihara,
and Y. Shimada, Enzymatic Enrichment of Astaxanthin from
Haematococcus pluvialis Cell Extracts, J. Am. Oil Chem. Soc.
80:975–981 (2003).
18. Haraldsson, G.G., A. Halldorsson, and E. Kulås, Chemoenzy-
matic Synthesis of Structured Triacylglycerols Containing Eicosa-
pentaenoic and Docosahexaenoic Acids, Ibid. 77:1139–1145
(2000).
19. Halldorsson, A., C.D. Magnusson, and G.G. Haraldsson,
Chemoenzymatic Synthesis of Structured Triacylglycerols by
Highly Regioselective Acylation, Tetrahedron 59:9101–9109
(2003).
20. Hoshino, T., T. Yamane, and S. Shimizu, Selective Hydrolysis
of Fish Oil by Lipase to Concentrate n-3 Polyunsaturated Fatty
Acids, Agric. Biol. Chem. 54:1459–1467 (1990).
21. McNeill, G.P., R.G. Ackman, and S.R. Moore, Lipase-Catalyzed
Enrichment of Long-Chain Polyunsaturated Fatty Acids, J. Am.
Oil Chem. Soc. 73:1403–1407 (1996).
22. Haraldsson, G.G., and B. Hjaltason, Fish Oils as Sources of Im-
portant Polyunsaturated Fatty Acids, in Structured and Modi-
fied Lipids, edited by F.D. Gunstone, Marcel Dekker, New
York, 2001, pp. 313–350.
23. Halldorsson, A., B. Kristinsson, and G.G. Haraldsson, Lipase
Selectivity Toward Fatty Acids Commonly Found in Fish Oil,
Eur. J. Lipid Sci. Technol. 106:79–87 (2004).
24. Lie, Ø., and G. Lambertsen, Digestive Lipolytic Enzymes in
Cod (Gadus morrhua): Fatty Specificity, Comp. Biochem.
Physiol. 80B:447–450 (1985).
REFERENCES
1. Matsuno, T., and S. Hirao, Marine Carotenoids, in Marine Bio-
genic Lipids, Fats, and Oils, edited by R.G. Ackman, CRC
Press, Boca Raton, FL, 1989, Vol. 1, pp. 251–388.
2. Miki, W., Biological Functions and Activities of Animal
Carotenoids, Pure Appl. Chem. 63:141–146 (1991).
3. Krinsky, N.I., The Biological Properties of Carotenoids, Ibid.
66:1003–1010 (1994).
4. Bjerkeng, B., M. Fölling, S. Lagocki, T. Storebakken, J.J. Olli,
and N. Alsted, Bioavailability of all-E-Astaxanthin and Z-
Isomers of Astaxanthin in Rainbow Trout (Oncorhynchus
mykiss), Aquaculture 157:63–82 (1997).
5. Torrissen, O.J., R.W. Hardy, and K.D. Shearer, Pigmentation
of Salmonids—Carotenoid Deposition and Metabolism, CRC
Crit. Rev. Aquat. Sci. 1:209–225 (1989).
6. Storebakken, T., and H.K. No, Pigmentation of Rainbow Trout,
Aquaculture 100:209–229 (1992).
7. Østerlie, M., B. Bjerkeng, and S. Liaaen-Jensen, Accumulation
of Astaxanthin all-E, 9Z and 13Z Geometrical Isomers and 3
and 3′ RS Optical Isomers in Rainbow Trout (Oncorhynchus
mykiss) Is Selective, J. Nutr. 129:391–398 (1999).
8. Johnson, E.A., and G.-H. An, Astaxanthin from Microbial
Sources, Crit. Rev. Biotechnol. 11:297–326 (1991).
9. Foss, P., T. Storebakken, E. Austreng, and S. Liaaen-Jensen,
Carotenoids in Diets for Salmonids. V. Pigmentation of Rain-
bow Trout and Sea Trout with Astaxanthin and Astaxanthin Di-
palmitate in Comparison with Canthaxanthin, Aquaculture
65:293–305 (1987).
10. Torrissen, O.J., Strategies for Salmonid Pigmentation, J. Appl.
Ichthyol. 11:276–281 (1995).
11. Aas, G.H., B. Bjerkeng, T. Storebakken, and B. Ruyter, Blood
Appearance, Metabolic Transformation and Plasma Transport
Proteins of 14C-Astaxanthin in Atlantic Salmon (Salmo salar
L.), Fish Physiol. Biochem. 21:325–334 (1999).
25. Gellesvik, D.R., Fatty Specificity of Bile-Salt Dependent Lipase:
Enzyme Recognition and Super-substrate Effects, Biochim. Bio-
phys. Acta 1086:167–172 (1991).
12. Arai, S., T. Mori, W. Miki, K. Yamaguchi, S. Konosu, M. Sa-
take, and T. Fujita, Pigmentation of Juvenile Coho Salmon with
[Received July 11, 2003; accepted February 16, 2004]
JAOCS, Vol. 81, no. 4 (2004)