Enzymatic carotenoid cleavage in star fruit (Averrhoa carambola)

Article abstract of DOI:10.1016/S0031-9422(02)00657-X

This paper presents the first description of an enzyme fraction exhibiting carotenoid cleavage activity isolated from fruit skin of Averrhoa carambola. Partial purification of the enzyme could be achieved by acetone precipitation, ultrafiltration (300 kDa

Full text of DOI:10.1016/S0031-9422(02)00657-X

Phytochemistry 63 (2003) 131–137
www.elsevier.com/locate/phytochem
Enzymatic carotenoid cleavage in star fruit (Averrhoa carambola)
a,
b
a
Peter Fleischmann *, Naoharu Watanabe , Peter Winterhalter
a
Institut fu¨r Lebensmittelchemie, Technische Universita¨t Braunschweig, Schleinitzstrasse 20, D-38106 Braunschweig, Germany
Department of Applied Biological Chemistry, Faculty of Agriculture, Shizuoka University, 836 Ohya, Shizuoka 422-4870, Japan
b
Received 29July 2002; received in revised form 5 November 2002
Abstract
This paper presents the first description of an enzyme fraction exhibiting carotenoid cleavage activity isolated from fruit skin of
Averrhoa carambola. Partial purification of the enzyme could be achieved by acetone precipitation, ultrafiltration (300 kDa, 50
kDa), isoelectric focusing (pH 3–10) and sodium dodecyl sulfate polyacrylamide gel electrophoresis (7.5%). In this way, an enzy-
matically active protein fraction was obtained, consisting of four proteins in the molecular weight range of between 12 and 90 kDa.
Using b-carotene as substrate, the enzyme activity was detected spectrophotometrically at 505 nm. The main reaction product,
detected by GC analysis, was b-ionone. This proves that the isolated enzymes are closely related to aroma metabolism and release
ꢀ
1
of star fruit. The time constant of the reaction was 16.6 min, the Michaelis Constant K =3.6 mmol 1 and the maximum velocity
m
ꢀ3
ꢀ1 ꢀ1
ꢀ1
ꢂ
V
max=10.5ꢁ10 mmol l
s
2003 Elsevier Science Ltd. All rights reserved.
mg(Protein) . The optimum temperature was 45 C.
#
Keywords: Averrhoa carambola; Fruit; Carotenoids; Oxidative cleavage; b-Ionone
1
. Introduction
cally degrading 9-cis-neoxanthin to abscisic acid
Schwartz et al., 1997), there is only information about
(
Volatile carotenoid breakdown products are long
known as important flavor compounds, in various fruit
Winterhalter and Rouseff, 2001a). The flavor of star
fruit (Averrhoa carambola) is strongly influenced by
these carotenoid derived compounds such as b-ionone
carotenoid cleavage enzymes in Arabidopsis thaliana
(Schwartz et al., 2001) and in quince (Fleischmann et
al., 2002), both with C13 and C14 compounds as most
likely reaction products.
In the present paper, we describe the partial purifica-
tion of an enzyme fraction with carotenoid cleavage
activity from star fruit (Averrhoa carambola, Oxalida-
ceae) skin. This cytosolic activity is characterized by its
(
(Winterhalter and Schreier, 1995). However, until now
the biochemical pathways giving rise to those flavor
compounds are not fully elucidated.
Apart from chemical oxidative degradation of car-
otenoids, enzymatic oxidative cleavage is seen as an
important flavor formation pathway (Enzell, 1985;
Wahlberg and Eklund, 1998; Winterhalter and Rouseff,
kinetic parameters (Vmax, K , time constant, tempera-
m
ture dependence) and its main reaction product.
2
001b; Winterhalter and Schreier, 1988). However,
2. Results and discussion
whereas enzymatic cleavage of carotenoids in animal
tissues is well known (Glover et al., 1960; Lakshman
and Mychkovsky, 1989; von Lintig and Vogt, 2000;
Wolf, 2001; Wyss et al., 2000), the number of reports
giving evidence for similar reactions in plant tissues is
still limited (Fleischmann et al., 2001, 2002; Schwartz et
al., 1997, 2001). With the exception of enzymes specifi-
2.1. Isolation and characterization of star fruit carotenoid
cleavage enzymes
Despite the highly lipophilic nature of their substrate,
carotenases are cytosolic proteins (Fleischmann et al.,
2001; Schwartz et al., 2001; von Lintig and Vogt, 2000).
Consequently, they are only found in water soluble frac-
tions of star fruit skins, whereas no activity could be traced
in other fractions or fruit tissues. Importantly, only
extracts from fully ripened star fruit showed carotenoid
*
Corresponding author. Tel.: +49-531-3917217; fax: +49-531-
917230.
E-mail address: p.fleischmann@tu-bs.de (P. Fleischmann).
3
0
031-9422/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0031-9422(02)00657-X

132
P. Fleischmann et al. / Phytochemistry 63 (2003) 131–137
ꢂ
Fig. 1. Kinetics of b-carotene degradation by enzymes isolated from unripe and fully ripened star fruit (Averrhoa carambola) (40 C, pH=7.0, 2.48
mmol/l initial b-carotene concentration, 24.1 mg/ml protein concentration). For ripe star fruit, an exponential regression curve was calculated (time
constant=16.6 min).
cleavage activity (Fig. 1). Extracts from unripe star fruit do
not show any measurable enzymatic carotenoid cleavage
activity. The slow decline in b-carotene concentration
shown in Fig. 1 is undistinguishable from chemical
degradation at the same conditions. After isoelectric
focusing, enzyme activity was concentrated in a single
peak (Fig. 2). The peak fractions were pooled as par-
tially purified carotenoid cleavage enzyme. After SDS
PAGE and silver staining, four protein bands were
detected in the enzyme solution (Fig. 3). The molecular
weight of the proteins was calculated from the calibra-
present at the SDS PAGE shown in Fig. 3. Together
with the fact that the two bands around 90 kDa and the
band at 12 kDa are present in inactive protein fractions
as well (lane 2, Fig. 3), it is most likely that the 50 kDa
band represents the enzyme responsible for the car-
otenoid cleavage described here. Furthermore, the
molecular weight of this protein is close to that reported
for carotenases isolated from other sources (66 kDa).
Carotenoid cleavage enzymes isolated from Arabidopsis
thaliana (Schwartz et al., 2001), as well as enzymes iso-
lated from chicken (related to retinal formation) (Wyss
et al., 2000) belong to this group. Additionally, VP14
type 9-cis-epoxy-carotenoid cleavage enzymes, which
are involved in abscisic acid formation in species like
tion curve of the standard proteins (R value vs. mol.
f
mass) used for the SDS PAGEs. Three bands between
5
0 and 90 kDa and one protein band at 12 kDa are
Fig. 2. Activity pattern of the enzyme after preparative isoelectric focusing. The peak shows the approximate value of the isoelectric point (pH=3.6)
of the enzyme.

P. Fleischmann et al. / Phytochemistry 63 (2003) 131–137
133
perature measured (Fig. 5). Despite the increasing non-
enzymatic b-carotene degradation at higher temperatures
(
Fig. 5), the data show that the optimal operating tem-
ꢂ
perature of this enzyme is approximately 45 C. This is
not surprising, because of the similar optimum tem-
perature of already known fruit carotenases found in
the same range (Fleischmann et al., 2001). The data
used for the determination of the optimum temperature
were subsequently used for the calculation of the acti-
vation energy E . Fig. 6 shows an Arrhenius plot of
a
these data in semi-logarithmic form (ln v!1/T). The
slope of the plot shown in Fig. 6 (calculated from the
equation of the regression curve) was used to calculate
ꢀ1
the activation energy E =101.5 kJ mol of the clea-
a
vage reaction. This value is higher than the typical acti-
vation energy levels of enzymes (Bisswanger, 1979). For
water soluble enzyme specifically oxidizing a highly
lipophilic substrate like b-carotene, enzymes and sub-
strate molecules (usually provided in micellar form)
should reside in different compartments of the incu-
bation mixture. Therefore it is not surprising to find
higher differences in energy levels (and therefore activa-
tion energy) between single substrate and enzyme mole-
cules on the one hand and the enzyme-substrate
complexes on the other.
Fig. 3. Polyacrylamide gel electrophoresis of the partially purified
enzyme.
avocado (Persea americana) (Chernys and Zeevaart,
2000), bean (Phaseolus vulgaris) (Qin and Zeevaart,
1999), and again Arabidopsis thaliana (Schwartz et al.,
1997), are known to have very similar molecular masses.
2
.2. Kinetic properties
The time constant of the reaction was calculated by
The optimum pH of b-carotene degradation activity
could be detected at pH 8.5 (Fig. 7). At pH values lower
than 7, Fig. 7 shows an increase in activity again. But
the blank data show that this is caused only by an
increase in non-enzymatic chemical degradation. These
results differ from those reported for enzymes derived
from quince, which did show activity changes dependent
on the pH of the incubation mixtures (Fleischmann et
al., 2002). In order to further investigate this dis-
crepancy, studies on the influence of the different origins
of the plants (quince from temperate European, star-
fruit from tropical climate) are currently planned.
non linear regression of the b-carotene degradation time
course (Fig. 1). The value obtained was 16.6 min. Fig. 1
shows that the time course of the carotenoid cleavage
reaction could be fitted to an exponential function, typi-
cal for first order reactions. Therefore, under our experi-
mental conditions, the influence of other substrates than
b-carotene (i.e. oxygen) is negligible. Consequently, it
was possible to use first order Michaelis–Menten kinetics
for a description of the carotenoid cleavage activity.
The kinetic parameters (Vmax and K ) of the car-
m
otenoid cleavage enzymes isolated from star fruit were
ꢂ
studied at pH 7.0 and 23 C. Km and Vmax were calcu-
lated by linear regression (R=0.99) of the enzyme’s
2.3. Reaction products
Lineweaver–Burk plot, using b-carotene as substrate
ꢀ
3
(
l
Fig. 4). The values obtained (V =10.5ꢁ10 mmol
Finally, GC experiments proved that the major vola-
tile reaction product isolated from incubations of
b-carotene with isolated star fruit carotenoid cleavage
enzymes is b-ionone. GC analyses of reaction products
extracted from active incubation mixtures as described
in 3.5 (attached to an SPME fiber) show a single peak at
35.8 min. Co-injection experiments of isolated reaction
products together with standard b-ionone also resulted
in a single peak (35.9min) under the same running
conditions in GC. Therefore we can conclude that
biooxidative carotenoid cleavage enzymes in star fruit
skins exclusively cleave b-carotene eccentrically at the
max
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
s
mg(Protein) ; K =3.6 mmol l ) are clearly dif-
ferent from values determined for plant carotenases
isolated from other sources (Fleischmann et al., 2001,
m
2
002). The Km of the star fruit carotenases described
here is approximately 4 times lower than the one repor-
ted for quince carotenases, whereas the Vmax is about 10
times higher compared with quince (Fleischmann et al.,
002). Therefore carotenases isolated from star fruit
skin show clearly higher affinity to the substrate b-car-
2
otene (K ) and also a higher turnover efficiency (Vmax)
m
compared to carotenases obtained from quince, sup-
porting the idea of at least two different types of car-
otenoid cleaving enzymes in fruit tissues.
The temperature profile of star fruit carotenases shows
an increase of enzyme activity until the maximum tem-
0
0
9–10 (9 –10 ) double bond with b-ionone as the dominant
volatile reaction product. Together with their dependence
on the ripening state of the fruit, these results prove the
importance of carotenoid cleavage enzymes for aroma

134
P. Fleischmann et al. / Phytochemistry 63 (2003) 131–137
Fig. 4. Lineweaver–Burk plot with regression curve and calculated kinetic key constants of star fruit carotenoid cleavage enzyme.
ꢂ
Fig. 5. Temperature dependence of star fruit cleavage enzymes. At 45 C, we could trace the highest enzyme activity. sd=Standard deviation.
production and metabolism in ripening and ripe star
fruit. However, until now no second reaction product,
i.e. carotenoid fragment, could be detected, what is
mainly due our methodology (GC analysis). Further
experiments are currently undertaken in order to iden-
tify non-volatile reaction products.
until fully ripened (at least 2 weeks). This ripening period
was needed to induce formation of intense flavor as well
as relevant amounts of carotenoid cleavage enzymes
inside the fruit-skin. During ripening, the green, almost
scentless fruit changed to intensively yellow color and
fruity, ionone-like aroma.
3.2. Chemicals
3
. Experimental
b-Carotene was purchased from Sigma (Japan). Bio-
lyte 3-10 ampholytes used for isoelectric focusing were
3
.1. Plant material
obtained from BioRad (Hercules CA), Tris–HCl,
MgCl , KCl, Tween 40, sodium dodecyl sulfate (SDS),
acrylamide, and bisacrylamide were all purchased from
Wako, Japan. All other reagents were at least of analy-
tical grade.
Star fruit (Averrhoa carambola) has been obtained
from a local fruit market in Shizuoka, Japan. In order to
isolate enzymatically active protein fractions, it was
2
ꢂ
necessary to store the fruit at room temperatures (24 C)

P. Fleischmann et al. / Phytochemistry 63 (2003) 131–137
135
ꢀ1
Fig. 6. Arrhenius plot of the star fruit cleavage enzymes. The activation energy of the cleavage reaction E =101.5 kJ mol could be determined
a
from the slope of the regression curve. The regression was performed for the black (filled) data points.
Fig. 7. pH dependence of star fruit carotenoid cleavage enzymes. The optimal pH could be found at pH=8.5. sd=Standard deviation.
3
.3. Enzyme isolation
to remove the ampholytes needed for isoelectric focus-
ing, the active fractions were ultrafiltrated at 50 kDa.
The protein fractions resulting from this step were
After removing the outer wax layer of fully ripened
star fruit (1.0 kg) with n-hexane (10 s), the peelings were
homogenized in sample buffer (125 mmol/l KCl, 5
mmol/l MgCl , 50 mmol/l Tris, pH 6.8, 100 ml per 100 g
ꢂ
stored in sample buffer (1 C) until used as enzyme
ꢂ
extract. All workup steps were carried out at 4 C.
2
of fresh fruit skin) using an ultra Waring blender (120
s). The homogenates were centrifuged at 2000 g for 10
min. The resulting pellet was discarded and the super-
3.4. Enzyme assay and kinetic methods
Carotenoid cleavage activity was measured spectro-
photometrically with b-carotene as substrate. b-Car-
otene (0.1 mg/ml) was dissolved in water using Tween
40 (10% in stock solution) as emulgator. The reaction
was carried out in microcuvettes (final volume 375 ml) in
the spectrophotometer, using an initial b-carotene con-
centration of 2.48 mmol/l in the reaction mixture. For
natant subjected to an acetone precipitation (87%, 3 h,
ꢂ
10.0 C). The resulting precipitate was dissolved in
sample buffer and subjected to a preparative isoelectric
ꢂ
focusing (Bio-Rad Rotofor, pH 3–10, 12 W, 7 h, 4 C, ).
All protein fractions resulting from isoelectric focusing
were screened for carotenoid cleavage activity. In order

136
P. Fleischmann et al. / Phytochemistry 63 (2003) 131–137
ꢂ
ꢂ
measuring the Lineweaver-Burk plot the assay was per-
formed at 9different substrate concentrations between
m; temperature profile: 35–230 C, 3 C/min; injection
ꢂ
temperature: 250 C; carrier gas: N , 1.0 ml min) after
2
0
.05 and 50 mmol/l. The enzymatic reaction was mea-
extraction from the reaction mixture. For extracting
volatile norisoprenoids from the incubated enzyme
solutions, Bio-Beads SM-2 (Bio Rad) were added
directly after the end of incubation. After 3 h, the Bio-
Beads were removed from the incubation mixture and
washed twice with Milli-Q for 10 s. The beads were then
transferred into a stirred methanol solution (20 ml, 2 h,
room temperature) to redissolve the volatiles. This step
was repeated three times. The methanol fractions were
combined and concentrated by a Vigreux column to a
final volume of 500 ml. Finally, the reaction products
were transferred to SPME fibers (Polymethylsiloxane,
sured against b-carotene control solutions under the
same reaction conditions, but without protein. Mole-
cular oxygen, the only co-substrate needed (and used)
for the carotenoid degradation reaction, was provided
in abundance using air saturated buffer solutions. The
carotenoid cleavage reactions were monitored con-
tinuously at 505 nm. At this wavelength, it was possible
to trace the b-carotene cleavage uninfluenced from the
spectral absorption of resulting reaction products.
The temperature dependency of enzyme activity was
measured similar to the method described above. How-
ever, here only the reaction cuvettes contained b-car-
otene. The reference cuvettes contained b-carotene free,
otherwise identical solutions. For the control measure-
ꢂ
100 mm; 10 min incubation time, 23 C), which were
directly used for the GC experiments.
ꢂ
ments, the enzymes were heat deactivated (90 C, 10
min) prior to the incubations. No difference in b-car-
otene degradation could be detected between protein
free incubation solutions and incubations containing
deactivated enzymes. The activation energy of the clea-
vage reaction was calculated using the logarithmic form
of the Arrhenius equation:
Acknowledgements
This work was supported by the Japanese Society for
the Promotion of Science (JSPS) and a Grant-in-Aid
from the Ministry of Education, Science, Sports and
Culture, Japan.
Ea
lnkcat ¼ lnk0 ꢀ
RꢃT
References
Bisswanger, H., 1979. Theorie und Methoden der Enzymkinetik. Ver-
lag Chemie, Weinhein, Basel, Deerfield Beach.
At the beginning of the incubation period, kcat can be
regarded as almost independent of the reverse reaction.
Consequently, for enzyme catalyzed reactions like the
cleavage of b-carotene, the ln kcat can be replaced by the
logarithm of reaction speed, ln v. The slope of a plot ln v
Chernys, J.T., Zeevaart, J.A.D., 2000. Characterization of the 9-cis-
epoxycarotenoid dioxygenase gene family and the regulation of
abscisic acid biosynthesis in avocado. Plant Physiol. 124, 343–353.
Enzell, C., 1985. Biodegradation of carotenoids—an important route
to aroma compounds. Pure and Appl. Chem. 57, 693–700.
Fleischmann, P., Lutz-Roeder, A., Winterhalter, P., Watanabe, N.,
2001. Carotenoid cleavage enzymes in animals and plants. In: Win-
terhalter, P., Rouseff, R. (Eds.), Carotenoid-Derived Aroma Com-
pounds. ACS Symp. Ser. 802, Vol. 219. American Chemical Society,
Washington, DC, pp. 76–88.
!
1/T, (multiplied by ꢀR) exhibits the activation energy
ꢀ1
E in J mol . The left part of the data set (Fig. 6) was not
a
included into the fit in order to avoid possible influence
of enzyme deactivation or heat denaturation on the cal-
culation of E . Finally it is necessary to note that the E
a
a
Fleischmann, P., Studer, K., Winterhalter, P., 2002. Partial purifica-
tion and kinetic characterization of a carotenoid cleavage enzyme
from quince fruit (Cydonia oblonga). J. Agric. Food Chem. 50,
1677–1680.
value calculated here is underestimating the real activa-
tion energy of this carotenase. This was unavoidable
because of the extremely high absorbance of b-carotene
at saturating conditions (50 mmol/l; A>3).
Glover, J., Deretic, D., Papermaster, D.S., 1960. The conversion of
b-carotene into vitamin A. Vitamins and Hormones 18, 371–386.
Lakshman, M.R., Mychkovsky, I., 1989. Enzymatic conversion of all-
trans-b-carotene to retinal by a cytosolic enzyme from rabbit and
rat intestinal mucosa. Proc. Natl. Acad. Sci. USA 86, 9124–9128.
Qin, X.Q., Zeevaart, J.A.D., 1999. The 9-cis-epoxycarotenoid clea-
vage reaction is the key regulatory step of abscisic acid biosynthesis
in water-stressed bean. Proc. Natl. Acad. Sci. USA 96, 15354–
3
.5. Determination of proteins and reaction products
The protein content was measured according to the
Biuret method using BSA as standard protein.
Denaturing polyacrylamide gel electrophoresis (SDS-
PAGE) was carried out in a Mini Protein Cell (Wako
Corp.) using commercially available 7.5% poly-
acrylamide gels (Wako Corp.). The gels were stained
with a commercially available silver stain kit (Wako
Corp. Japan). All runs were carried out under constant
voltage conditions (90 V) at room temperature.
1
5361.
Schwartz, S.H., Qin, X.Q., Zeevaart, J.A.D., 2001. Characterization of
a novel carotenoid cleavage dioxygenase from plants. J. Biol. Chem.
276, 25208–25211.
Schwartz, S.H., Tan, B.C., Gage, D.A., Zeevaart, J.A.D., McCarty,
D.R., 1997. Specific oxidative cleavage of carotenoids by VP14 of
maize. Science 276, 1872–1874.
von Lintig, J., Vogt, K., 2000. Filling the gap in vitamin A research—
molecular identification of an enzyme cleaving beta-carotene to ret-
inal. J. Biol. Chem. 275, 11915–11920.
The reaction products were detected by gas chroma-
tography (Column: GL Sciences TC-WAX, 0.25 mm, 30

P. Fleischmann et al. / Phytochemistry 63 (2003) 131–137
137
Wahlberg, I., Eklund, A.-M., 1998. Degraded carotenoids. In: Britton,
G., Liaaen-Jensen, S., Pfander, H. (Eds.), Carotenoids, Vol. 3: Bio-
in quince (Cydonia oblonga, Mill.) fruit. J. Agric. Food Chem. 36,
1251–1256.
synthesis and Metabolism. Birkha
¨
user, Basel, pp. 195–216.
Winterhalter, P., Schreier, P., 1995. The generation of norisoprenoid
volatiles in starfruit (Averrhoa carambola L.): a review. Food Rev.
Int. 11, 237–254.
Winterhalter, P., Rouseff, R. (Eds.), 2001a. Carotenoid-Derived
Aroma Compounds. ACS Symp. Ser. 802, Vol. 219. American
Chemical Society, Washington, DC.
Wolf, G., 2001. The enzymatic cleavage of b-carotene: end of a con-
troversy. Nutrition Reviews 59, 116–118.
Winterhalter, P., Rouseff, R., 2001b. Carotenoid-derived aroma com-
pounds: An introduction. In: Winterhalter, P., Rouseff, R. (Eds.),
Carotenoid-Derived Aroma Compounds. ACS Symp. Ser. 802, Vol.
Wyss, A., Wirtz, G., Woggon, W.D., Brugger, R., Wyss, M.,
Friedlein, A., Bachmann, H., Hunziker, W., 2000. Cloning and
0
219. American Chemical Society, Washington, DC, pp. 1–17.
expression of b, b-carotene 15,15 -dioxygenase. Biochem. Biophys.
Res. Commun. 271, 334–336.
Winterhalter, P., Schreier, P., 1988. Free and bound C13 norisoprenoids