An NADPH and FAD dependent enzyme catalyzes hydroxylation of flavonoids in position 8

Article abstract of DOI:10.1016/j.phytochem.2006.03.008

Yellow flavonols contribute to flower pigmentation in Asteraceae. In contrast to common flavonols, they show additional hydroxyl groups in position 6 and/or 8 of the aromatic A-ring in addition to the basic 5,7-hydroxylation pattern. An enzyme introducing a hydroxyl group in position 8 of flavonols and flavones was demonstrated for the first time with enzyme preparations from petals of Chrysanthemum segetum. Flavanones, dihydroflavonols and glucosylated flavonols and flavones were not accepted as substrates. The enzyme was localized in the microsomal fraction and uses NADPH and FAD as cofactors. Experiments with carbon monoxide/blue light and with antibodies specific for cytochrome P450 reductase did not indicate the involvement of a classical cytochrome P450 dependent monooxygenase in the reaction. Thus, the flavonoid 8-hydroxylase represents a novel type of hydroxylating enzyme in the flavonoid pathway. Apart from flavonoid 8-hydroxylase activity, the presence of all enzymes involved in the formation of flavonoid 7-O-glucosides in C. segetum was demonstrated. The pathway leading to 8-hydroxyflavonoids in C. segetum has been derived from enzyme activities and substrate specificities observed.

Full text of DOI:10.1016/j.phytochem.2006.03.008

PHYTOCHEMISTRY
Phytochemistry 67 (2006) 1080–1087
www.elsevier.com/locate/phytochem
An NADPH and FAD dependent enzyme catalyzes
hydroxylation of flavonoids in position 8
Heidrun Halbwirth, Karl Stich *
Technische Universit a¨ t Wien, Institut f u¨ r Verfahrenstechnik, Umwelttechnik und Technische Biowissenschaften,
Getreidemarkt 9/1665, A-1060 Wien, Austria
Received 2 December 2005; received in revised form 1 March 2006
Available online 6 May 2006
Abstract
Yellow flavonols contribute to flower pigmentation in Asteraceae. In contrast to common flavonols, they show additional hydroxyl
groups in position 6 and/or 8 of the aromatic A-ring in addition to the basic 5,7-hydroxylation pattern. An enzyme introducing a hydro-
xyl group in position 8 of flavonols and flavones was demonstrated for the first time with enzyme preparations from petals of Chrysan-
themum segetum. Flavanones, dihydroflavonols and glucosylated flavonols and flavones were not accepted as substrates. The enzyme was
localized in the microsomal fraction and uses NADPH and FAD as cofactors. Experiments with carbon monoxide/blue light and with
antibodies specific for cytochrome P450 reductase did not indicate the involvement of a classical cytochrome P450 dependent monooxy-
genase in the reaction. Thus, the flavonoid 8-hydroxylase represents a novel type of hydroxylating enzyme in the flavonoid pathway.
Apart from flavonoid 8-hydroxylase activity, the presence of all enzymes involved in the formation of flavonoid 7-O-glucosides in C.
segetum was demonstrated. The pathway leading to 8-hydroxyflavonoids in C. segetum has been derived from enzyme activities and sub-
strate specificities observed.
Ó 2006 Elsevier Ltd. All rights reserved.
Keywords: Chrysanthemum segetum; Asteraceae; Flavonoids; A-Ring hydroxylation; Flavonoid 8-hydroxylase; FAD; Higher hydroxylated flavonols;
Gossypetin; 8-Hydroxyluteolin
1
. Introduction
tin), gossypetin (8-hydroxyquercetin) and their derivatives
owe their color to the presence of an extra hydroxyl group
in position 6 or 8 of the aromatic A-ring (Fig. 1) (Har-
borne, 1967). Whereas members of common flavonols are
wide-spread and may be found in nearly all plant families,
the occurrence of such higher hydroxylated flavonols seems
to be much more restricted, but their methylated and/or
glucosylated derivatives occur in Asteraceae, Malvaceae
and Leguminosae (Harborne, 1967). In flowers of Astera-
ceae, 6-hydroxyflavonols are part of the pigmentation
pattern fulfilling important physiological functions as
honey guides (Thompson et al., 1972; Halbwirth et al.,
2000). In addition, the main pigments of cotton flowers are
based on gossypetin, and gossypetin derivatives also con-
tribute to yellow flower color in Chrysanthemum segetum,
Lotus corniculatus, Hibiscus tiliaecus and Primula vulgaris
(Geissman and Steelink, 1957; Harborne, 1967).
Flavonoids have a common, basic C6–C3–C6 skeleton
structure consisting of two aromatic rings (A and B) and
a heterocyclic ring (C) containing one oxygen atom
(
Fig. 1). The most common flavonoids show a basic
0
0 0
5
,7,4 - or a 5,7,3 4 -hydroxylation pattern. Furthermore,
most flavonoid classes possess hydroxyl groups in ring C.
Yellow flavonols such as quercetagetin (6-hydroxyquerce-
Abbreviations: CHI, chalcone isomerase; CHS, chalcone synthase;
0
DFR, dihydroflavonol 4-reductase; FHT, flavanone 3-hydroxylase; F3 H,
flavonoid 3 -hydroxylase; F8H, flavonoid 8-hydroxylase; F6H, flavonol 6-
hydroxylase; F7GT, Flavonoid 7-O-glucosyltransferase; FNS II, flavone
synthase II.
0
*
Corresponding author. Tel.: +43 1 58801 17320; fax: +43 1 58801
17399.
E-mail address: kstich@mail.zserv.tuwien.ac.at (K. Stich).
0
031-9422/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2006.03.008

H. Halbwirth, K. Stich / Phytochemistry 67 (2006) 1080–1087
1081
OH
3
´
OH
R1
8
4´
B
HO
R2
7
O
2
A
C
6
3
R
5
4
OH
O
R=OH, R
R=OH, R
R=OH, R
1
1
1
=H,
=OH, R
=H,
2
R =H: 5,7,3´,4´-Tetrahydroxyflavonol (Quercetin)
2
=H:
5,7,8,3´,4´-Pentahydroxyflavonol (8-Hydroxyquercetin)
R
2
=OH: 5,6,7,3´,4´-Pentahydroxyflavonol (Quercetagetin)
R=H,
R=H,
R=H,
R
R
R
1
=H,
1
=OH, R
1
=H,
R
2
=H: 5,7,3´,4´-Tetrahydroxyflavone (Luteolin)
=H: 5,7,8,3´,4´-Pentahydroxyflavone (8-Hydroxyluteolin)
2
=OH: 5,6,7,3´,4´-Pentahydroxyflavone (6-Hydroxyluteolin)
2
R
Fig. 1. Chemical structures of flavonoids showing additional hydroxyl groups in ring A.
In the flavonoid pathway, hydroxyl groups in the A-, B-
2. Results and discussion
and C-rings are introduced at various biosynthetic stages.
As a rule, the enzymes catalyzing these reactions are either
2.1. Introduction of the hydroxyl group in position 8
2
3
-oxoglutarate-dependent dioxygenases such as flavanone
-hydroxylase (FHT) or cytochrome P450 dependent mon-
1
4
Incubation of [ C]quercetin with crude enzyme prepara-
0
0
ooxygenases such as flavonoid 3 -hydroxylase (F3 H), fla-
tions from petals of C. segetum in the presence of NADPH
led to the formation of one [ C]-labeled product (Fig. 2A),
0
0
14
vonoid 3 ,5 -hydroxylase, and chalcone 3-hydroxylase
Forkmann and Heller, 1999; Wimmer et al., 1998). How-
(
which was identified as gossypetin by TLC with the authen-
tic reference substance in four different solvent systems.
Furthermore, the identity of the [ C]-labeled product was
confirmed by HPLC co-chromatography with the authentic
reference substance using a photodiode array detector cou-
pled with a radioactivity detector. In the presence of 2-oxo-
ever, particular hydroxyl groups can be also obtained by
reducing of oxo-groups (introduction of the hydroxyl
group in position 4) or by the action of a bifunctional poly-
phenol oxidase (hydroxylation during aurone formation)
1
4
(
2
Forkmann and Heller, 1999; Strack and Schliemann,
001).
During the past few years, hydroxylation in position 6
2
+
glutarate, ascorbate and Fe , no formation of gossypetin
could be observed. The enzyme seems to be membrane-
bound, since a more than 2-fold enrichment of F8H activity
was obtained in the microsomal protein fraction compared
to the crude extract, whereas only 10% of the activity
remained in the supernatant after the final precipitation of
the microsomal fraction (Table 1). No enzyme activity
was observed, when preparations from leaves instead of pet-
als were used. Therefore, all further enzymatic studies were
performed with the microsomal fraction from petals. To
investigate the potential involvement of oxygen in the reac-
tion, an oxygen consuming system consisting of glucose,
glucose oxidase and catalase was used according to Kochs
and Grisebach (1987) (Table 2). Interestingly, exclusion of
oxygen did not lead to a total loss of enzymatic activity as
described for other hydroxylating enzymes (Wimmer
et al., 1998; Halbwirth et al., 2004). However, the most
striking observation during these investigations was the dis-
tinct enhancement of the enzyme activity (almost 3-fold
activity compared to the control) in the presence of heat-
inactivated catalase and glucose oxidase (Table 2). One pos-
sible explanation was the release of FAD from the active
center of glucose oxidase after enzyme denaturation during
boiling. Therefore, we studied the involvement of further
potential cofactors in the reaction.
has been studied in more detail. Interestingly, two differ-
ent enzyme systems were shown to catalyze this reaction
in the flavonoid pathway. Anzellotti and Ibrahim (2000,
2
004) were able to show a novel 2-oxoglutarate-dependent
dioxygenase, flavonol 6-hydroxylase (F6H), to be respon-
sible for the 6-hydroxylation of the partially methylated
flavonols in Chrysosplenium americanum leaves. In con-
trast, we have previously shown another F6H of petals
of Tagetes species, which could be classified as cyto-
chrome P450 dependent monooxygenase (Halbwirth
et al., 2004), and which is specifically involved in flower
pigment formation. Furthermore, Latunde-Dada et al.
(
6
2001) reported a cytochrome P450 dependent flavonoid
-hydroxylase from elicitor treated soybean, which seems
to be specifically involved in the biosynthesis of 6-
hydroxyisoflavones.
Hydroxylation in position 8 had not been studied
before, but a flavonol 7-O-glucosyltransferase (F7GT) cata-
lyzing the glucosylation of gossypetin in position 7 was
already characterized in detail from petals of C. segetum
(
Stich et al., 1997). We report on the occurrence of a micro-
somal oxidoreductase dependent on NADPH and FAD in
the petals of C. segetum catalyzing the 8-hydroxylation of
flavonols and flavones with very high efficiency and high
specificity.
Addition of FAD up to a concentration of 0.01 mM led
to a distinct increase in product formation (Fig. 3), but

1082
H. Halbwirth, K. Stich / Phytochemistry 67 (2006) 1080–1087
Fig. 2. Radioscan of TLC on cellulose from incubation of enzyme preparations from C. segetum in the presence of NADPH and FAD with (A)
14
14
[
C]quercetin as a substrate (solvent system 1) (B) [ C]luteolin as a substrate (solvent system 2).
Table 1
Table 2
Dependence on cofactors and subcellular localization of the flavonoid 8-
hydroxylase
Dependence of the flavonoid 8-hydroxylase on oxygen
Additions
Relative activity (%)
Enzyme source
Additions
Relative
activity (%)
480 lM NADPH
100
48
5
0 mM glucose, 5 U glucose oxidase,
0 U catalase
Crude extract
Crude extract
None
0
0
1
500 lM 2-oxoglutarate,
5
5
0 mM glucose
0 mM glucose, 5U boiled glucose oxidase,
101
280
100 lM FeSO
4
Crude extract
480 lM NADPH
100
10
220
50
1
0 U boiled catalase
Supernatant
480 lM NADPH
4
4
80 lM NADPH, 0.1 lM FMN
80 lM NADPH, 0.1 lM Haem
98
91
Microsomal fraction
Microsomal fraction
Microsomal fraction
Microsomal fraction
Microsomal fraction
480 lM NADPH
480 lM NADH
480 lM NADPH, 0.1 lM FAD
480 lM NADH, 0.1 lM FAD
0.1 lM/100 mM FAD
498
51
0/0
All activities were expressed per kg of total protein and relative activities
were calculated compared to the activity of the microsomal fraction
obtained in the presence of 480 lM NADPH. 100% correspond to
1
.3 lkat/kg protein.
All activities were expressed per kg of total protein and relative activities
were calculated compared to the activity obtained with crude extract in the
presence of 480 lM NADPH. 100% correspond to 1.4 lkat/kg protein.
The effect of various potential cofactors was tested using microsomal
preparations.
nounced loss of enzyme activity of approximately 75%.
In the presence of NADH, however, addition of FAD
did not result in enhanced enzyme activities (Table 1).
Thus, FAD and NADPH were added as cofactors for all
further studies. The subcellular localization in the micro-
somal fraction and the dependence on oxygen could also
be confirmed in assays containing FAD and NADPH as
cofactors (data not shown).
stimulated the reaction only in the presence of NADPH
(
Table 1). FMN and Haem (prosthetic group of catalase)
were not able to affect the reaction in the same way (Table
). Substitution of NADPH by NADH led to a pro-
2

H. Halbwirth, K. Stich / Phytochemistry 67 (2006) 1080–1087
1083
5
.5
4
4
3
2
1
0
.5
3
.5
2
.5
1
.5
0
0
0.02
0.04
0.06
FAD [mM]
0.08
0.1
Fig. 3. Influence of FAD on flavonoid 8-hydroxylase activity.
Beside quercetin, the flavone luteolin was converted to
a great extent to the corresponding 8-hydroxyluteolin
time for up to 30 min and with protein concentration up
to 2.5 lg protein in the standard assay. At 5 °C and
40 °C, the rate was approximately 50% of the maximum.
Studies on temperature stability revealed that pre-incuba-
tion of the enzyme solution for 10 min at up to 30 °C
had no influence on enzyme activity. Further increases in
temperature led to a marked reduction of enzyme activity.
When the microsomal fraction was frozen in liquid nitro-
gen and stored at ꢀ80 °C, no loss of activity could be
(
Fig. 2B). Thus, the F8H shows a broader substrate
specificity than the F6H from Tagetes sp., which clearly
preferred flavonols and hydroxylated flavones only to a
very small extent (Halbwirth et al., 2000). When
kaempferol was used as a substrate, no 8-hydroxyka-
empferol could be detected, but the formation of gossy-
petin was observed. In case of apigenin as substrate,
only 8-hydroxyluteolin and small amounts of luteolin
were formed. This may be easily explained by the high
observed over two years. The values for apparent K and
m
Vmax were 9 lM and 133 lkat/kg protein for quercetin
0
F3 H activity, which is also present in the microsomal
(Vmax/K = 14.8) and 10 lM and 72 lkat/kg for luteolin
m
preparations of C. segetum petals. No reaction products
were detected, when quercetin 3-O-glucoside, quercetin
(Vmax/K = 7.2), respectively. Since Vmax/Km provides an
m
indication of substrate specificity, quercetin seems to be
the preferred substrate for F8H. The F8H activity was
studied during the development of buds and flowers
(Fig. 4). The highest specific activities were measured in
enzyme preparations from closed buds and buds just open-
ing (stages 1–3), where the specific activity was 2.9 lkat/kg
protein under standard conditions. However, relatively
high activities could be still observed in open flowers
(stages 4 and 5, Fig. 4).
The data obtained raised the question, whether F8H is a
cytochrome P450 dependent monooxygenase as many
other hydroxylating enzymes in the flavonoid pathway
(Forkmann and Heller, 1999). These enzymes are mem-
7
-O-glucoside or luteolin 7-O-glucoside were incubated
under the same conditions. Flavanones, dihydroflavonols
0
and 6 -deoxychalcones were not accepted as substrates
either.
The reaction showed an optimum at pH 7.0 for the
hydroxylation of quercetin and at pH 6.75 for the hydrox-
ylation of luteolin. Fifty percent of the maximum were still
observed at pH 6.0 and 8.5, respectively. Thus, the reaction
takes place within a rather broad pH range. A more
detailed characterization was performed with quercetin as
a substrate. Highest reaction rates were measured at
3
0 °C, where the formation of gossypetin was linear with
3
2
1
0
.5
3
.5
2
.5
1
.5
0
1
2
3
4
5
6
Stage of flower development
Fig. 4. Dependence of flavonoid 8-hydroxylase activity on the stage of flower development.

1084
H. Halbwirth, K. Stich / Phytochemistry 67 (2006) 1080–1087
brane-bound haem proteins with a minimal catalytic system
constituted from the cytochrome P450 and the membrane-
bound flavoprotein NADPH-cytochrome P450-reductase
to a moderate loss of F8H activity, whereas the cytochrome
P450 dependent monooxygenase F3 H, which was also
0
present in the same preparations, was drastically affected.
A control assay with flavanone 3-hydroxylase (FHT)
revealed no influence of the antibodies on the 2-oxoglutar-
ate dependent dioxygenase (Table 4). These results clearly
indicate that the enzyme does not belong to the classical
cytochrome P450 dependent monooxygenases commonly
found in the flavonoid pathway. However, the nature of
the enzyme still remains unclear. There are many reports
on FAD dependent monooxygenases (Wagner et al.,
1990; Eppink et al., 2000; Luying and Sandvik, 2000).
Some are even localized in the microsomal fraction (Tynes
and Hodgson, 1984; Nagase et al., 1992), but most of them
have been found in micro-organisms and they seem to be
involved in catabolic processes rather than in the biosyn-
thesis of secondary compounds. In contrast, the enzyme
from C. segetum seems to be specifically involved in the
8-hydroxylation of flavonols and flavones.
(
Bolwell et al., 1994). Subcellular localization in the micro-
somal fraction and NADPH dependence are typical for this
enzyme class. The use of FAD is rather unusual, but could
be explained if the NADPH-cytochrome P450-reductase
from C. segetum easily looses FAD. However, in contrast
to F8H, well-known cytochrome P450 dependent monoox-
0
ygenases from the flavonoid pathway such as F3 H and
FNS II were not activated by addition of FAD. Therefore,
we investigated the effects of potential inhibitors on the
enzymatic formation of gossypetin (Table 3). Addition of
EDTA and cytochrome c did not influence the enzyme reac-
0
tion, whereas p-hydroxymercuribenzoate, 2,2 -dipyridyl or
N-ethyl maleimide were able to inhibit the reaction to a
moderate extent. The cytochrome P450 specific inhibitors
ancymidol and BAS 111 W had only a weak influence,
whereas the cytochrome P450 specific inhibitors ketocona-
zole and tetcyclacis clearly reduced enzyme activity. How-
ever, the influences observed were rather moderate
compared to other cytochrome P450 dependent monooxy-
genases (Stich et al., 1988; Halbwirth et al., 2004; Wimmer
et al., 1998).
The most specific reactions for the identification of cyto-
chrome P450 monooxygenases are the photoreversible inhi-
bition of carbon monoxide and the addition of cytochrome
P450 reductase specific antibodies (Benveniste et al., 1989;
Nielsen and Møller, 1999; Halbwirth et al., 2004). The
hydroxylation reaction was partially inhibited when the
microsomal preparation was pre-incubated with carbon
monoxide (CO) (Table 3). Pre-incubation with nitrogen
did not have a comparable effect. However, the CO inhibi-
tion could not be reversed by exposition of the enzyme
assay to blue light (Table 3). The addition of a crude frac-
tion of cytochrome P450 reductase specific antibodies led
The fact that the cytochrome P450 reductase specific
0
antibodies (Benveniste et al., 1989) inhibited F3 H but
0
not F8H, allowed us to study whether 4 -hydroxylated
flavonoids are accepted as substrates by F8H. In the pres-
ence of the antibodies, the formation of 8-hydroxykaempf-
erol from kaempferol and 8-hydroxyapigenin from
0
0
apigenin was observed. Thus, a 3 ,4 -hydroxylation pattern
is not a precondition for the introduction of an additional
hydroxyl group in position 8.
2.2. Biochemical pathway leading to gossypetin 7-O-
glucoside in C. segetum
Besides F8H activity, the presence of all enzymes
involved in the formation of flavonoid 7-O-glucosides
could be demonstrated in preparations from C. segetum
petals: Chalcone synthase/chalcone isomerase (CHS/
CHI), flavone synthase II (FNS II), FHT, flavonol syn-
Table 3
0
Effect of CO and other potential inhibitors on the flavonoid 8-hydroxylase
activity from C. segetum
thase (FLS), F3 H and flavonoid 7-O-glucosyltransferase
F7GT). Dihydroflavonol 4-reductase (DFR) activity
(
Assay condition
Relative activity (%)
could not be detected, thus providing an explanation for
the absence of anthocyanins in the yellow cultivar Helios.
The pathway leading to gossypetin 7-O-glucoside is shown
in Fig. 5. Chalcone synthase (CHS) clearly preferred p-cou-
maroyl-Co A as a substrate. Activities were only two times
higher than those observed with caffeoyl-CoA, but when
p-coumaroyl-CoA and caffeoyl-CoA were present in the
assay in equimolar concentrations, 90% naringenin
Control
Control with blue light
CO
100
98
76
73
101
99
CO with blue light
N
N
2
2
with blue light
EDTA (5 mM)
pOHMB (0.1 mM)
100
31
(
derived from p-coumaroyl-CoA) and 10% eriodictyol
N-Ethyl maleimide (5 mM)
53
67
100
90
30
0
0
2,2 -Dipyridyl (1 mM)
(derived from caffeoyl-CoA) were formed. F3 H showed
very high activity converting naringenin, apigenin,
dihydrokaempferol and kaempferol to the respective
Cytochrome c (0.5 mM)
Ancymidol (0.05 mM)
Ketoconazole (0.05 mM)
Tetcyclacis (0.05 mM)
BAS 111 W (0.05 mM)
0
0
0
0
3
,4 -hydroxylated products. Thus, the 3 ,4 -hydroxylation
58
87
pattern seems to be introduced at the flavonoid level by
F3 H rather than by CHS using caffeoyl-CoA as a sub-
strate. Since flavanones and dihydroflavonols were not
hydroxylated in position 8, it is obvious that the formation
0
Relative activities were calculated compared to a control. 100% corre-
spond to 4 lkat/kg protein. The data represent an average of three inde-
pendent experiments.

H. Halbwirth, K. Stich / Phytochemistry 67 (2006) 1080–1087
1085
Table 4
Influence of cytochrome P450 reductase specific antibodies on selected enzymes of the flavonoid pathway
Enzyme tested
Type of enzyme
Relative activities (%) in the presence of 10 lg
Cyt. P450 red. specific antibodies
Boiled Cyt. P450 red. specific antibodies
F8H
F3 H
FHT
NADPH, FAD dependent oxidoreductase
Cytochrome P450 dependent monooxygenase
2-Oxoglutarate dependent dioxygenase
85
0
100
100
100
100
0
Enzyme activities were measured in the presence of 10 lg cytochrome P450 reductase specific antibodies. Relative activities were calculated compared to a
control.
0
0
Abbreviations. F8H: Flavonoid 8-hydroxylase, F3 H: Flavonoid 3 -hydroxylase, FHT: Flavanone 3-hydroxylase, cyt.: cytochrome, red.: reductase.
of flavonols and flavones has to precede A-ring hydroxyl-
ation. However the hydroxyl group is introduced at the
aglycone level, since flavone and flavonol glycosides are
not accepted as substrates. This corresponds well with the
substrate preference of F7GT for gossypetin (Stich et al.,
1997) in C. segetum.
OH
OH
OH
OH
HO
O
O
-
-
3
HOOC CH2 COSCoA
+ ACoS
Malonyl-CoA
OH
O
p-Coumaroyl-CoA
8-Hydroxyluteolin
CHS/CHI
OH
OH
F8H
OH
OH
OH
HO
O
O
HO
O
O
HO
O
FNSII
F3'H
OH
OH
O
OH
Naringenin
FHT
Luteolin
Eriodictyol
FHT
OH
OH
OH
HO
O
HO
O
F8H
F3'H
OH
OH
OH
O
OH
O
Dihydrokaempferol
Dihydroquercetin
OH
OH
FLS
FLS
OH
OH
OH
OH
HO
O
HO
O
HO
O
O
F3'H
F8H
OH
OH
OH
OH
O
OH
O
OH
Gossypetin
Quercetin
Kaempferol
F8H
OH
OH
F7GT
F7GT
OH
OH
F8H GluO
OH
OH
OH
HO
O
GluO
O
O
O
OH
OH
OH
OH
O
OH
OH
O
Gossyspetin 7-O-glucoside
Quercetin 7-O-glucoside
8-Hydroxykaempferol
Fig. 5. Biochemical pathway leading to gossypetin 7-O-glucoside in C. segetum. (Bold arrows indicate putative main routes derived from enzyme activities
and substrate specificities observed).

1086
H. Halbwirth, K. Stich / Phytochemistry 67 (2006) 1080–1087
It may be assumed that 8-hydroxyluteolin is formed also
came from W. Heller (GSF, Munich, Germany). NADPH,
cytochrome c, glucose oxidase and catalase were obtained
from Sigma (Vienna, Austria). Cytochrome P450 specific
in vivo in the C. segetum petals, since high FNS II activity
was observed and luteolin represents an excellent substrate
for the F8H. However, the presence of 8-hydroxyluteolin in
C. segetum has not been reported so far although the flavo-
noids in this plant have been investigated in detail (Geiss-
man and Steelink, 1957). Possible explanations could be
the preference of F8H for quercetin and the competition
of FHT and FNS II for common substrates. However, pos-
sible involvement of multi-enzyme complexes and substrate
channelling should also be considered. Further work on
this topic is in progress.
inhibitors were
Germany).
a gift from BASF (Limburgerhof,
3.4. Enzyme preparations and protein determination
Enzyme preparations from petals were obtained as
described (Halbwirth et al., 2002), using 0.1 M KH PO /
2
4
K HPO (containing 0.4% ascorbate, pH 7.0). Microsomal
2
4
preparations from C. segetum and D. variabilis were
obtained according to Wimmer et al. (1998). The micro-
somal fraction was resuspended in buffer, shock frozen in
liquid nitrogen and stored at ꢀ80 °C. Protein content was
determined by a modified Lowry procedure (Sandermann
and Strominger, 1972) using crystalline bovine serum albu-
min as a standard.
3
. Experimental
3
.1. General experimental procedures
TLC was performed on Merck pre-coated cellulose
(
without fluorescence indicator, 1.0571.001, Merck,
3
.5. Enzyme characterization
Darmstadt, Germany). Radiolabeled substances were
detected with a TLC Linear Analyzer (Berthold LB 2842,
Wildbad, Germany). HPLC analysis was performed on a
Perkin–Elmer Series 200 system equipped with a photodi-
ode array detector coupled with a 500TR Flow Scintilla-
tion Analyzer for the detection of radiolabeled substances
All data represent an average of at least three indepen-
dent experiments. Subcellular determination was carried
out according to Halbwirth et al. (2000). Determination
of the pH optimum was carried out as described for the
standard assay, but using 0.2 M KH PO –K HPO buffer
(
Perkin–Elmer, Vienna, Austria).
2
4
2
4
(containing 0.4% sodium ascorbate) with pH values
between 6.0 and 8.5. Exclusion of oxygen from the enzyme
assay was carried out as described (Wimmer et al., 1998)
using an oxygen consuming system consisting of glucose,
glucose oxidase and catalase. Carbon monoxide inhibition
assays were performed according to Nielsen and Møller
3
.2. Plant material
The investigations were performed on the commercial
variety ‘Helios’ (Meisert Samenzucht, Hannover, Ger-
many) of C. segetum. The plant material was cultivated
in the municipal parks of Vienna and collected during the
summer periods 1997–2000. Dahlia variabilis cv. Rubens
was purchased from Wirth (Vienna, Austria).
(1999) with the exception that a gas mixture composed of
carbon monoxide/oxygen (9:1, v/v) was used. Kinetic data
were calculated from Lineweaver–Burk plots. Determina-
tion of the apparent Michaelis constant (K ) and maxi-
m
mum reaction velocity (Vmax
performed with fixed concentrations of 480 lM NADPH
and 0.01 lM FAD.
)
for quercetin were
3
.3. Chemicals
1
4
UDP-D-[U- C]Glucose
(12.0 GBq/mmol)
and
1
4
[
2- C]malonyl-coenzyme A (2 GBq/mmol) were puchased
1
4
from Amersham International (UK). [ C]Naringenin,
3.6. Standard enzyme assay
1
4
14
14
[
[
C]eriodictyol, [ C]kaempferol, [ C]dihydrokaempferol,
C]dihydroquercetin, [ C]quercetin and [ C]quercetin
1
4
14
14
The reaction mixture contained in a final volume of
100 ll: 2 ll microsomal preparation (2.5 lg protein), 88 ll
0.1 M KH PO /K HPO (containing 0.4% ascorbate pH
1
4
7
-O-glucoside and [ C]quercetin 3-O-glucoside were pre-
pared as previously described (Halbwirth et al., 2004,
2
4
2
4
1
4
14
14
2
006). [ C]Luteolin was synthesized from [ C]naringenin
7.0), 0.048 nmol [ C]quercetin (108 Bq), 0.001 nmol FAD
using microsomal preparations from D. variabilis cv.
Rubens for flavone formation and introduction of a 3 -
and 48 nmol NADPH (each dissolved in 5 ll H O). The
2
0
reaction was started by addition of NADPH. After
15 min at 30 °C the reaction was stopped with 10 ll glacial
acetic acid. The phenolic compounds were extracted twice
with EtOAc (70 + 50 ll, respectively), the organic phases
were applied to a precoated cellulose plate and chromato-
graphed using solvent system 1. Radioactivity was detected
and quantified by scanning the plates with a TLC-linear
analyzer (Berthold, Wildbad, Germany).
hydroxyl group (Stich et al., 1988). The cytochrome
P450 reductase specific antiserum from Jerusalem arti-
choke was a gift from F. Durst and I. Benveniste (CNRS,
Strasbourg, France). Quercetin, quercetagetin, kaempf-
erol, apigenin, luteolin and eriodictyol were purchased
from Extrasynthese (Genay, France), gossypetin from
Apin Chemicals (Abingdon, UK). 8-Hydroxyluteolin

H. Halbwirth, K. Stich / Phytochemistry 67 (2006) 1080–1087
1087
3
.7. Stages of flower development
Halbwirth, H., Wurst, F., Forkmann, G., Stich, K., 2000. Biosynthesis of
yellow flower pigments belonging to flavonoids and related com-
pounds. In: Mohan, R.M. (Ed.), Research Advances in Phytochem-
istry, vol. 1. Global Research Network, Kerala, pp. 35–49.
Morphological criteria were used for subdividing the
developmental process into six different stages. Stages 1
and 2 refer to closed buds of 5 and 7 mm length, respec-
tively, stage 3 are buds just opening with a petal length
7
3
Halbwirth, H., Kampan, W., Stich, K., Fischer, T.C., Meisel, B.,
Forkmann, G., Rademacher, W., 2002. Biochemical and molecular
biological investigations with respect to induction of fire blight
resistance in apple and pear by transiently altering the flavonoid
metabolism with specific enzyme inhibitors. Acta Hort. 590, 485–492.
Halbwirth, H., Forkmann, G., Stich, K., 2004. The A-ring specific
hydroxylation of flavonols in position 6 in Tagetes sp. is catalyzed by a
cytochrome P450 dependent monooxygenase. Plant Sci. 167, 129–135.
Halbwirth, H., Kahl, S., J a¨ ger, W., Reznicek, G., Forkmann, G.,
mm. Stages 4 and 5 represent open flowers 25 and
5 mm in diameter, respectively, and stage 6 wilting flowers.
3
.8. Chromatography
14
Stich, K., 2006. Synthesis of ( C)-labeled 5-deoxyflavonoids and
their application in the study of dihydroflavonol/leucoanthocyani-
din interconversion by dihydroflavonol 4-reductase. Plant Sci. 170,
Substrates and products were separated and identified
by TLC on precoated cellulose using the following solvent
system: (1) n-butanol/acetic acid/water (6:1:2, v/v/v); (2)
chloroform/acetic acid/water (10:9:1, v/v/v); (3) t-buta-
nol/acetic acid/water (3:1:1,v/v/v); (4) 30% acetic acid.
Product identification was performed by HPLC analysis
according to Vande Casteele et al. (1982) (R quercetin:
2
min; R 8-hydroxyluteolin: 22.71 min).
5
87–595.
Harborne, J.B., 1967. Comparative Biochemistry of the Flavonoids.
Academic Press, London.
Kochs, G., Grisebach, H., 1987. Induction and characterization of a
NADPH-dependent flavone synthase from cell cultures of soybean. J.
Biosci. 42, 343–348.
Latunde-Dada, A.O., Cabello-Hurtado, F., Czittrich, N., Didierjean, L.,
Schopfer, C., Hertkorn, N., Werck-Reichhart, D., Ebel, J., 2001.
Flavonoid 6-hydroxylase from soybean (Glycine max L.), a novel plant
P450 monooxygenase. J. Biol. Chem. 276, 1688–1695.
t
4.66 min; R gossypetin: 21.62 min; R luteolin: 25.74
t t
t
Luying, X., Sandvik, E.R., 2000. Characterization of 4-hydroxyphenyl-
acetate 3-hydroxylase (HpaB) of Escherichia coli as a reduced flavin
adenine dinucleotide-utilizing monooxygenase. Appl. Environ. Micro-
biol. 66, 481–486.
Nagase, S., Aoyagi, K., Sakamoto, M., Takemura, K., Ishikawa, T.,
Narita, M., 1992. Biosynthesis of methylguanidine in the hepatic
microsomal fraction. Nephron 62, 182–186.
Nielsen, J.S., Møller, B.L., 1999. Biosynthesis of cyanogenic glucosides in
Triglochin maritima and the involvement of cytochrome P450 enzymes.
Arch. Biochem. Biophys. 368, 121–130.
Sandermann, H., Strominger, J.L., 1972. Purification and properties of
C55-isoprenoid alcohol phosphokinase from Staphylococcus aureus. J.
Biol. Chem. 247, 5123–5131.
Acknowledgements
These investigations were supported by a grant from the
Austrian FWF (Project 13020). We are grateful to F. Durst
and I. Benveniste for their gift of cytochrome P450
reductase specific antiserum. We would like to thank W.
Rademacher for gifts of various cytochrome P450 specific
inhibitors. Finally, we are indebted to E. Meggeneder for
critically reading the manuscript.
Stich, K., Ebermann, R., Forkmann, G., 1988. Effect of cytochrome P-450
References
0
specific inhibitors on the activity of flavonoid-3 -hydroxylase and
flavone synthase II in certain plants. Phyton 28, 237–247.
Stich, K., Halbwirth, H., Wurst, F., Forkmann, G., 1997. UDP-glucose:
flavonol 7-O-glucosyltransferase activity in flower extracts of Chry-
santhemum segetum. Z. Naturforsch c 52, 153–158.
Anzellotti, D., Ibrahim, R.K., 2000. Novel flavonol 2-oxoglutarate
dependent dioxygenase: Affinity purification, characterization, and
kinetic properties. Arch. Biochem. Biophys. 382, 161–172.
Anzellotti, D., Ibrahim, R.K., 2004. Molecular characterization and
functional expression of flavonol 6-hydroxylase. BMC Plant Biol. 4,
Strack, D., Schliemann, W., 2001. Bifunctional polyphenol oxidases: novel
functions in plant pigment biosynthesis. Angew. Chem. Int. Ed. 40,
3791–3794.
20.
Benveniste, I., Lesot, A., Hasenfratz, M.P., Durst, F., 1989. Immuno-
chemical characterization of NADPH-cytochrome P450 reductase
from Jerusalem artichoke and other higher plants. Biochem. J. 259,
Thompson, W.R., Meinwald, J., Aneshansley, D., Eisner, T., 1972.
Flavonols: responsible for ultraviolet absorption in nectar guide of
flower. Science 177, 528–530.
8
47–853.
Tynes, R.E., Hodgson, E., 1984. The measurement of FAD-containing
mono-oxygenase activity in microsomes containing cytochrome P-450.
Xenobiotica 14, 515–520.
Bolwell, P.G., Bozak, K., Zimmerlin, A., 1994. Review article number 96:
Plant cytochrome P450. Phytochemistry 37, 1491–1506.
Eppink, M.H.M., Cammaart, E., Van Wassenaar, D., Middelhoven, W.J.,
Van Berkel, W.J.H., 2000. Purification and properties of hydroquinone
Vande Casteele, K., Geiger, H., Van Sumere, C.F., 1982. Separation of
flavonoids by reversed-phase high-performance liquid chromatogra-
phy. J. Chromatogr. 240, 81–94.
hydroxylase,
a FAD-dependent monooxygenase involved in the
catabolism of 4-hydroxybenzoate in Candida parapsilosis CBS604.
Eur. J. Biochem. 267, 6832–6840.
Wagner, E.D., Verdier, M.M., Plewa, M.J., 1990. The biochemical
mechanisms of the plant activation of promutagenic aromatic amines.
Env. Mol. Mutagen. 15, 236–244.
Forkmann, G., Heller, W., 1999. Biosynthesis of flavonoids. In: Barton,
D., Nakanishi, K., Meth-Cohn, O. (Eds.), Comprehensive Natural
Products Chemistry, vol. 2. Elsevier, Oxford, pp. 713–748.
Geissman, T.A., Steelink, C., 1957. Flavonoid petal constituents of
Chrysanthemum segetum L. J. Org. Chem. 22, 946–948.
Wimmer, G., Halbwirth, H., Wurst, F., Forkmann, G., Stich, K.,
0
1998. Enzymatic hydroxylation of 6 -deoxychalcones with protein
preparations from petals of Dahlia variabilis. Phytochemistry 47,
1013–1016.