A non-canonical caleosin from Arabidopsis efficiently epoxidizes physiological unsaturated fatty acids with complete stereoselectivity

Article abstract of DOI:10.1111/j.1742-4658.2012.08757.x

In plants, epoxygenated fatty acids (EFAs) are constituents of oil seeds as well as defence molecules and components of biopolymers (cutin, suberin). While the pleiotropic biological activities of mammalian EFAs have been well documented, there is a paucity of information on the physiological relevance of plant EFAs and their biosynthesis. Potential candidates for EFA formation are caleosin-type peroxygenases which catalyze the epoxidation of unsaturated fatty acids in the presence of hydroperoxides as co-oxidants. However, the caleosins characterized so far, which are mostly localized in seeds, are poor epoxidases. In sharp contrast, quantitative RT-PCR analysis revealed that PXG4, a class II caleosin gene, is expressed in roots, stems, leaves and flowers of Arabidopsis. Expressed in yeast, PXG4 encodes a calcium-dependent membrane-associated hemoprotein able to catalyze typical peroxygenase reactions. Moreover, we show here that purified recombinant PXG4 is an efficient fatty acid epoxygenase, catalyzing the oxidation of cis double bonds of unsaturated fatty acids. Physiological linoleic and linolenic acids proved to be the preferred substrates for PXG4; they are oxidized into the different positional isomers of the monoepoxides and into diepoxides. An important regioselectivity was observed; the C-12,13 double bond of these unsaturated fatty acids being the least favored unsaturation epoxidized by PXG4, linolenic acid preferentially yielded the 9,10-15,16-diepoxide. Remarkably, PXG4 catalyzes exclusively the formation of (R),(S)-epoxide enantiomers, which is the absolute stereochemistry of the epoxides found in planta. These findings pave the way for the study of the functional role of EFAs and caleosins in plants. 2012 The Authors Journal compilation

Full text of DOI:10.1111/j.1742-4658.2012.08757.x

A non-canonical caleosin from Arabidopsis efficiently
epoxidizes physiological unsaturated fatty acids with
complete stereoselectivity
1
1
1
2
´
ˆ
Elizabeth Blee , Martine Flenet , Benoıt Boachon and Marie-Laure Fauconnier
´ ´
1 Institut de Biologie Moleculaire des Plantes, Universite de Strasbourg, France
´ `
2 Plant Biology Unit, Gembloux Agro-Bio Tech, Universite de Liege, Belgium
Keywords
In plants, epoxygenated fatty acids (EFAs) are constituents of oil seeds as
well as defence molecules and components of biopolymers (cutin, suberin).
While the pleiotropic biological activities of mammalian EFAs have been
well documented, there is a paucity of information on the physiological
relevance of plant EFAs and their biosynthesis. Potential candidates for
EFA formation are caleosin-type peroxygenases which catalyze the epoxi-
dation of unsaturated fatty acids in the presence of hydroperoxides as
co-oxidants. However, the caleosins characterized so far, which are mostly
localized in seeds, are poor epoxidases. In sharp contrast, quantitative
RT-PCR analysis revealed that PXG4, a class II caleosin gene, is expressed
in roots, stems, leaves and flowers of Arabidopsis. Expressed in yeast,
PXG4 encodes a calcium-dependent membrane-associated hemoprotein
able to catalyze typical peroxygenase reactions. Moreover, we show here
that purified recombinant PXG4 is an efficient fatty acid epoxygenase, cat-
alyzing the oxidation of cis double bonds of unsaturated fatty acids. Physi-
ological linoleic and linolenic acids proved to be the preferred substrates
for PXG4; they are oxidized into the different positional isomers of the
monoepoxides and into diepoxides. An important regioselectivity was
observed; the C-12,13 double bond of these unsaturated fatty acids being
the least favored unsaturation epoxidized by PXG4, linolenic acid preferen-
tially yielded the 9,10-15,16-diepoxide. Remarkably, PXG4 catalyzes exclu-
sively the formation of (R),(S)-epoxide enantiomers, which is the absolute
stereochemistry of the epoxides found in planta. These findings pave the
way for the study of the functional role of EFAs and caleosins in plants.
AtClO4; caleosin; epoxygenase; fatty acid
epoxide; peroxygenase
Correspondence
´
´
E. Blee, Institut de Biologie Moleculaire des
Plantes, 28 Rue Goethe, 67083 Strasbourg
Cedex, France
Fax: +33 3 68 85 19 21
Tel: +33 3 68 85 24 36
E-mail: Elizabeth.Blee@ibmp-cnrs@unistra.fr
(Received 7 June 2012, revised 15 August
2012, accepted 16 August 2012)
doi:10.1111/j.1742-4658.2012.08757.x
Introduction
Among mammalian oxylipins, epoxy fatty acids func-
tion as important biomediators in inflammation, blood
pressure regulation, angiogenesis and numerous physi-
ological and pathophysiological processes in the
central nervous system (reviewed recently in [1–3]).
The biosynthesis of these metabolites involves various
cytochrome P450 isoforms that oxidize unsaturated
fatty acids into a multitude of bioactive epoxygenated
fatty acid (EFA) regioisomers [4–7].
In comparison, very little is known about the
biosynthesis and the biological roles of EFAs in
plants. These oxylipins are classically found in seed
oils as constituents of triglycerides that accumulate in
lipid droplets (LDs). Despite the fact that the different
Abbreviations
13 HOT, 13-hydroxy-9,11,15-octadecatrienoic acid; 13 HPOD, 13-hydroperoxy-9,11-octadecadienoic acid; 13 HPOT, 13-hydroperoxy-9,11,15-
octadecatrienoic acid; EFA, epoxygenated fatty acid; EH, epoxide hydrolase; LD, lipid droplet.
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
1

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
positional isomers of EFAs deriving from unsaturated
fatty acids such as 9,10-, 12,13- and 15,16-epoxy deriv-
atives are present in such oils [8], the main focus
remained on the formation of vernolic acid, a particu-
lar epoxide deriving from linoleic acid. Found at high
amounts in the seeds of some exotic plants, vernolic
acid (12,13-epoxy-cis-9-octadecenoic acid) has several
potential industrial applications. It is used, for exam-
ple, in the production of plastics, adhesives, nylon and
paints [9,10]. Considerable effort has therefore been
devoted to the identification and expression of the
genes controlling its biosynthesis in commercial oil
crops. Depending on the species two unrelated epoxy-
genases have been identified: a delta 12 non-heme
diiron protein in the genus Crepis [11–13] and a cyto-
chrome P450 in the genus Euphorbia [14,15]. However,
because of the high regioselectivity of these two
enzymes which is restricted to the oxidation of the
C-12,13 double bond of linole(n)ic acid, the identity of
the biosynthetic enzymes responsible for the occur-
rence of the numerous other regioisomer and stereoiso-
mer epoxides present in the triglycerides of seed LDs
remain to be established.
In rice, it was shown that epoxy derivatives of lino-
leic and linolenic acids arise from a distinct branch of
the lipoxygenase pathway encompassing an uncharac-
terized
lipid-hydroperoxide-decomposing
activity
[26,27]. In parallel, in vitro studies have revealed the
presence, in various seeds and young seedlings, of
hydroperoxide-dependent peroxygenase activities that
potently catalyze the epoxidation of unsaturated fatty
acids [28,29]. Such peroxygenases were also suggested
to be involved in the biosynthesis of C18 cutin mono-
mers [30,31]. Peroxygenase activities were recently
demonstrated to be associated with caleosins, a small
family of proteins produced by two classes of genes in
plants [32]. Class I genes encode for the few caleosins
that have been biochemically characterized so far. But
these caleosins when heterologously expressed showed
only marginal epoxidase activity under standard condi-
tions [32,33]. Therefore class I caleosins can hardly
account for an efficient fatty acid epoxidation of
unsaturated fatty acids in green plants.
At present, no class II caleosins, named AtClO4 to
AtClO7 (Fig. 1) in Arabidopsis [34], have been charac-
terized biochemically. AtClO5 (At1g23240) was
detected in pollen coat and according to data from the
public microarray database its gene is exclusively
expressed in flowers, suggesting a role for this caleosin
in flowering [35]. In sharp contrast, AtClO4, though
present in flowers, is also expressed in roots, leaves
and stems, suggesting that AtClO4 might be function-
ally active not only in reproductive systems but also in
germination and in vegetative processes [36]. That class
II caleosins are peroxygenases, or even enzymes,
remains, however, to be demonstrated.
EFAs are not confined to the seeds. They are also
components of cutin, the scaffold of cuticles that cover
all the vegetative parts of plants [16]. Cutin monomers
were as well found to act as endogenous stress-related
signal molecules [17] inducing, for example, resistance
of barley against Erysiphe graminis [18]. Although
most cutin monomers originate from saturated or
mono-unsaturated fatty acids, oxygenated derivatives
of linoleic and linolenic acids such as 9,10-, 12,13- and
15,16-epoxides have also been reported, albeit more
rarely. For example they have been detected in rice
species resistant to Magnaporthe grisae and have been
found to accumulate as defence compounds in sensi-
tive rice species infected by this fungus [19,20]. Gener-
ally, EFAs do not accumulate in cells due to their
toxicity and their rapid hydrolysis by efficient epoxide
hydrolases (EHs). Selective inhibition of EHs in mam-
malian systems has allowed the modulation of EFA
levels and the unravelling of their biological functions
(recently reviewed in 21,22). Several inhibitors of
mammalian EHs were also shown to inhibit plant EHs
in vitro [23–25]; however, their selectivity, mode of
administration and bioavailability have not yet been
studied in planta. Alternatively, to assess the biological
significance of EFAs in plants the modulation of their
cellular levels by altering the gene expression of the
epoxidases responsible for their biosynthesis might also
be envisaged. Unfortunately, this genetic approach is
currently hampered by the lack of genes known to
encode such epoxygenases.
Class I
At5g55240
At5g29560
At2g33380
PXG1
At1g23240
At1g70680
PXG4
At1g23250
Class II
Fig. 1. Phylogenic tree of caleosins putatively present in
A. thaliana with At4g26740 as PXG1 and At1g70670 as PXG4.
2
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
Currently, we do not know which particular
enzymes are responsible for the biosynthesis of plant
EFAs. We expect candidates for this type of activity
to be part of the caleosin-type peroxygenases which
were shown to catalyze epoxidation reactions in
various seedlings. Here we report that AtClO4 encodes
a non-seed-specific caleosin able to efficiently catalyze
the epoxidation of unsaturated fatty acid double
bonds. We found that this epoxidation represents,
relative to other types of oxidation, the major reaction
catalyzed by AtClO4 (renamed here PXG4). Impor-
tantly these epoxidations, which are characterized by a
very high stereoselectivity, give access to the stereois-
omers found in planta. We therefore suggest that
AtClO4 is a gene implicated in fatty acid epoxidation
in Arabidopsis thaliana.
in the western blot analysis of the soluble extract
(Fig. 2A) but was no longer visible after its filtration
on a 0.75 lm diameter filter or its centrifugation at
high speed for 2 h. Therefore, PXG4 and class I caleo-
sins are mostly membrane-bound proteins that can be
recruited to LDs.
We recently reported that PXG1 like other class I
caleosins supports peroxygenase activity, i.e. it cata-
lyzed co-oxidation reactions, such as sulfoxidation,
hydroxylation or epoxidation, in the presence of
hydroperoxides as co-oxidants [32]. Crude extracts of
yeast harboring PXG1 efficiently oxidized aniline but
very poorly epoxidized unsaturated fatty acids such as
oleic acid (Fig. 2B). Routinely, we measured a ratio of
1 : 0.3 between the specific activity of aniline hydroxyl-
ation and oleic acid epoxidation respectively. In con-
trast, [14C]oleic acid was rapidly metabolized by PXG4
in a strictly hydroperoxide-dependent manner. Up to
19 nmolÁminÀ1ÁmgÀ1 protein of cis-9,10-epoxystearate
was formed; this was about 20-fold more than the
production of hydroxylated aniline by the same yeast
crude extract (Fig. 2B). Importantly, radio-labeled
epoxy derivative was the sole metabolic product
detectable during the oxidation of [14C]oleic acid by
PXG4 in the presence of cumene hydroperoxide. Nota-
bly, no [14C]dihydroxy-octadecanoic acid could be
detected confirming that the yeast used for expressing
caleosins had no EH activity [38].
Results
Characterization of PXG4 as a membrane-bound
fatty acid epoxygenase
Class II caleosins lack the canonical central hydropho-
bic domain that is present in class I caleosins and is
presumably responsible for their LD targeting/docking
[37]. To determine whether class II members are still
membrane-bound proteins, we expressed PXG1
(At4g26740) and PXG4 (At1g70670) in yeast as repre-
sentative of classes I and II respectively. We then
assessed the ability of the recombinant proteins to be
located on LDs. Western blot analysis of crude
extracts of the transformed yeast revealed that, com-
pared with PXG1 expression, PXG4 was far more
substantially expressed (about fourfold, Fig. 2A). Both
isoforms were actively targeted to LDs. The excess of
overexpressed PXG4 was apparently retained in the
microsomal fraction containing the endoplasmic reticu-
lum (Fig. 2A). A thinner band of PXG4 was present
Oleic acid exhibits a cis double bond at position
9,10. Keeping the unsaturation at the same position
while shortening the length of the alkyl chain did not
noticeably affect the epoxidation rates (Table 1). In
contrast, shifting the unsaturation of octadecenoic acid
to position 6 or 11 favored the epoxidation of the
double bond closest to the carboxylate (Table 1). In
addition, PXG4 exhibited a low selectivity with respect
to the configuration of the double bond. For example,
cis and trans vaccenic acids or oleic and elaidic acids
A
B
30 kDa
24 kDa
20
10
0
PXG4
PXG1
PXG1
PXG4
PXG1
PXG4
LD
M
S
Fig. 2. (A) Western blot analysis of crude yeast extracts expressing PXG1 and PXG4 using anti-His6 mouse monoclonal IgG: LD, lipid
droplet; M, microsomes; S, supernatant. (B) Specific activities of PXG1 and PXG4 measured as aniline hydroxylases (black striped) and oleic
acid epoxygenases (black). Enzymatic activities were corrected by the respective expression levels of the two caleosins.
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
3

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
Table 1. Substrate specificity of PXG4. Mono-unsaturated and
polyunsaturated fatty acids (200 lM) were incubated in the
presence of purified PXG4 (2 lg) and cumene hydroperoxide
(200 lM) for 30 min at 27 °C. The resulting epoxides and remaining
substrates were extracted with ether, methylated with ethereal
diazomethane and analyzed by gas chromatography. The given
results represent mean values SE of three biological replicates.
indicating that a free carboxylic group seems necessary
for the positioning of the substrate within the active
site. Overall these results indicate that, in contrast to
the class I caleosins characterized to date, PXG4
is an effective hydroperoxide-dependent fatty acid
epoxygenase.
Unsaturated fatty acids
Reaction progress (%)
Purification and characterization of recombinant
PXG4
C14:1-9 cis
23
36
32
15
22
53
31
12
64
74
11
1
2
2
1
1
2
2
1
3
4
1
C16:1-9 cis
C18:1-9 cis
To further characterize the catalytic activity of PXG4
we purified a His6-tagged version of the recombinant
enzyme on an Ni2+ column. The purified protein gave
C18:1-9 trans
C20:1-11 cis
C18:1-6 cis
a
single band on SDS/PAGE at 24 2 kDa as
C18:1-11 cis
revealed by Coomassie Blue staining and western blot
analysis using an anti-His-tag IgG (Fig. 3A). In addi-
tion, a minor protein band with a molecular mass of
about 48 kDa was apparent, suggesting that PXG4
might also exist as a dimer, like the class I caleosins.
The term caleosin partly originates from the presence
of a strictly conserved EF-hand calcium binding motif
that was shown, at least for PXG1, to be crucial for
its catalytic activity [32]. Because the EF-hand motif
present in PXG4 typifies it as a caleosin, we verified
C18:1-11 trans
C18:2-9,12 cis
C18:3-9,12,15 cis
C18:1-9 cis methyl ester
were both epoxidized by this enzyme, albeit at a lower
rate for the trans- compared with the cis-isomer
(Table 1). It should be noted here that the methyl ester
of oleic acid was a poor substrate for PXG4 (Table 1)
A
B
a
b
c
80
60
40
20
0
43
26
11
Fig. 3. (A) Analysis of purified
recombinant PXG4 by western blot using
1
2
3
(a) anti-His-tag IgG, (b) Coomassie Blue
staining and (c) molecular weight
0.04
C
0.002
standards. (B) Purified PXG4 actively
catalyzed aniline hydroxylation in the
presence of cumene hydroperoxide (1).
Extensive dialysis against EDTA
completely abolished this activity (2),
which could be restored (3) by addition of
0.02
0
500
2 mM CaCl2. These experiments were
600
λ (nm)
repeated twice with different batches of
purified PXG4 with similar results. (C)
Absolute absorption spectra (solid line)
and difference spectrum (dithionite
reduced minus oxidized, dashed line) of
PXG4. Inset, difference spectrum
(dithionite reduced minus oxidized) in
the presence of 20% pyridine (v/v) in
0.1 M NaOH.
0
350
450
550
Wavelength (nm)
650
750
4
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
whether the activity of this protein was also calcium-
dependent. Extensive dialysis of purified PXG4 against
EDTA totally abolished the capacity of this fraction
to oxidize aniline in the presence of cumene hydroper-
oxide as co-oxidant. This enzymatic activity could be
restored, however, by addition of CaCl2 (Fig. 3B).
These results indicate that calcium was also required
for the structure/activity of PXG4.
linoleate and linolenate are of great interest, since their
oxidized derivatives have been found to be involved in
plant defence mechanisms or in jasmonate production
[19,39]. Increasing the desaturation of C18 fatty acids
seemed to greatly facilitate their epoxidation (Table 1).
We have confirmed this result by incubating [1-14C]li-
noleic and [1-14C]linolenic acids in the presence of
cumene hydroperoxide and PXG4. Epoxidation
occurred readily, i.e. under similar experimental condi-
tions the rate of substrate disappearance (expressed as
V/Km) was about sevenfold to tenfold higher for lino-
lenic and linoleic acids respectively than for oleic acid
(Fig. 4A). Two groups of compounds were formed:
the monoepoxides (peaks 2 and 4 in Fig. 4A) and the
diepoxides (peaks 3 and 5 in Fig. 4A). The rate of for-
mation of the monoepoxides from C18:2 by PXG4
was about two orders of magnitude faster than the
formation of the diepoxide. Curiously, no triepoxide
arising from linolenic acid oxidation was detected
(Fig. 4Ac). In addition, when the monoepoxides deriv-
ing from C18:3 oxidation (Fig. 4A, peak 4) were puri-
fied from the TLC plate and further incubated with
PXG4 in the presence of cumene hydroperoxide, only
diepoxides accumulated (not shown). To determine
whether such results stem from a possible regioselectiv-
ity of PXG4, we separated the different monoepoxide
and diepoxide isomers by straight phase (SP) HPLC
using a silica column under isocratic conditions. The
two monoepoxides of C18:2 were baseline separated
(Fig. 4Bd) and identified by their fragmentation pat-
tern using MS. A slight but constant preference for the
formation of 9,10-epoxy-12-octadecenoic acid (60%
2%; Fig. 4B, peak 7) was observed during the linoleic
acid epoxidation by PXG4. Notably, the ratio between
the two regioisomers did not change until the reaction
was completed. In addition, the oxidation of the cis
double bond at position 9,10 prevailed over that at
positions 12,13 and 15,16 when linolenic acid was used
as the substrate, i.e. 9,10-epoxy-12,15-octadecadienoic
acid (Fig. 4B, peak 10) represented one-third of the
monoepoxides formed whereas similar amounts of
15,16-epoxy-9,12-octadecadienoic acid (26%; Fig. 4B,
peak 9) and 12,13-epoxy-9,15-octadecadienoic acid
(24%; Fig. 4B, peak 8) accumulated. The regioselectiv-
ity of PXG4 appeared far more pronounced in
diepoxide formation. A major peak of 9,10-15,16-diep-
oxy-12-octadecenoic acid (76%; Fig. 4B, peak 11)
eluted initially on the silica column of SP-HPLC
followed by a small peak of 9,10-12,13 diepoxy-15-oc-
tadecenoate (22%; Fig. 4B, peak 12) and a very small
peak of 12,13-15,16-diepoxy-9-octadecenoate (2%;
Fig. 4B, peak 13). Taken together, these data indicate
that PXG4 is a regioselective enzyme. Epoxidation of
Both the absolute spectrum of oxidized versus dithi-
onite reduced PXG4 (Fig. 3C) and the formation of a
pyridine–ferrohemoprotein complex in the presence of
pyridine and NaOH (Fig. 3C, inset) at 556 nm sug-
gested that PXG4 contained a protoheme of type b
(0.6 nmol of heme per mg protein has been estimated).
Indeed, the electronic absorption spectrum of PXG4
(Fig. 3C) is characteristic of a high-spin ferric heme
site, with the Soret band at 407 nm and a charge
transfer band at 635 nm (Fig. 3C). Such a spectrum is
very similar to that of soybean peroxygenase [28] and
PXG1 [32]. The EPR spectrum of PXG1 showed that
the heme site is in the ferric high-spin oxidation state
that is pentacoordinated with histidine as an axial
ligand [32]. Hence, the heme site in resting (as isolated)
PXG4 is concluded to be probably a pentacoordinated
heme in the ferric oxidation state. Such configuration
is favorable for the epoxidation reaction with hydro-
peroxides as oxidants (to form the oxoferryl intermedi-
ate state) and lipids as substrates. The prosthetic heme
is most probably not covalently bound to the protein
because dialysis of purified PXG4 against a mildly
acidic buffer (pH 5.5) led to its progressive loss, as
monitored by the severe decrease in intensity of the
Soret band at 407 nm, yielding finally the apoprotein
devoid of catalytic activity. Site-directed mutagenesis
has indicated His70 and His133 as the putative heme
proximal and distal ligands in PXG1 [32]. Similarly to
PXG1, replacing the putative axial ligand (His90 in
PXG4 and His70 in PXG1) resulted in complete
impairment of the enzymatic activity of PXG4, indi-
cating that this residue is probably required for the
structure and/or activity of this protein. Considering
the quite favorable yield of expression of the mem-
brane-bound PXG4 in yeast [5–7 (mg protein)ÁLÀ1 of
culture], we shall try to crystallize this LD protein.
Indeed, X-ray determination should help to rationalize
site-directed mutagenesis approaches to identify the
active site residues of PXG4.
PXG4 is a regioselective peroxygenase
PXG4 catalyzed the epoxidation of mono-unsaturated
fatty acids, such as oleic acid. From a physiological
point of view polyunsaturated fatty acids such as
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
5

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
A
B
1000
500
1
a
Relative V/K
0.13
d
7
6
C18:1
0
1500
0
100
e
f
10
b
c
3
1
9
8
2
0
9000
C18:2
11
0
100
5
0.7
4
12
13
C18:3
0
0
0
30
0
20
Migration (cm)
Time (min)
Fig. 4. (A) Radio-TLC analysis of the products resulting from the incubation of (a) oleic acid (C18:1), (b) linoleic acid (C18:2) and (c) linolenic
acid (C18:3) by PXG4 in the presence of cumene hydroperoxide. Peak 1, 9,10-epoxy-stearic acid; peak 2, 9,10-12,13-diepoxy-stearic acid;
peak 3, mixture of 9,10-epoxy and 12,13-epoxy oleic acids; peak 4, mixture of 9,10-epoxy, 12,13-epoxy and 15-16-epoxy linoleic acids; peak
5, mixture of 9,10,12,13-diepoxy, 9,10,15,16-diepoxy and 12,13,15,16-diepoxy oleic acids. (B) Radio-SP-HPLC analysis of the monoepoxides
resulting from incubation of C18:2 (d) and the monoepoxides (e) and diepoxides (f) resulting from incubation of C18:3 with PXG4 in the
presence of cumene hydroperoxide. Peak 6, 12,13-epoxy-9-octadecenoic acid; peak 7, 9,10-epoxy-12-octadecenoic acid; peak 8, 12,13-
epoxy-9,15-octadecadienoic acid; peak 9, 15,16-epoxy-9,12-octadecadienoic acid; peak 10, 9,10-epoxy-12,15-octadecadienoic acid; peak 11,
9,10-15,16-diepoxy-12-octadecenoic acid; peak 12: 9,10-12,13-diepoxy-15-octadecenoic acid; peak 13, 12,13-15,16-diepoxy-9-octadecenoic
acid.
polyunsaturated fatty acids was taking place preferen-
tially at position 9,10, then at 15,16 and lastly at
12,13.
In addition, when linoleic acid was used as the sub-
strate instead of oleic acid, PXG4 produced only the
enantiomer (R),(S) of each regioisomer. Therefore, and
in sharp contrast with all the peroxygenases studied to
date [28,29,40], PXG4 formed only the 9(R),10(S)-
epoxy (Fig. 5Bhi) and the 12(R),13(S)-epoxy deriva-
tives of linoleic acid (Fig. 5Bjk). We can conclude
from these experiments that PXG4 catalyzed the total
asymmetric epoxidation of unsaturated fatty acids. It
is worth noting, however, that when the methyl ester
of oleic acid was used instead of the free fatty acid,
the enantiomeric excess of the resulting epoxide
dropped from > 99% to 40% (Fig. 5Bfg).
PXG4 exhibits a complete stereoselectivity
forming only (R),(S) enantiomers
We have previously reported that PXG1 catalyzed
asymmetric oxidation [32]. For example, when oleic
acid was used as the substrate, PXG1 produced the
cis-9,10-epoxy-stearic acid with the preferential forma-
tion of 9(R),10(S) enantiomer (40% ee). To evaluate
whether PXG4 was also a stereospecific enzyme, we
incubated [14C]oleic acid in the presence of PXG4 and
cumene hydroperoxide, H2O2, 13-hydroperoxy-9,11
-octadecadienoic acid (13 HPOD) or 13-hydroper-
oxy-9,11,15-octadecatrienoic acid (13 HPOT) as
co-oxidants. After the reaction was completed, the
enantiomeric composition of the resulting epoxides
was determined by chiral phase HPLC. Remarkably,
only one epoxide enantiomer (cis-9(R),10(S)-epoxy-
stearic acid) was formed (> 99% ee), whatever hydro-
peroxides had been used as co-oxidant (Fig. 5Aa–e).
PXG4 catalyzed intermolecular and
intramolecular transfer of oxygen
Using [14C]13 HPOT as mechanistic probe [41], we
have previously demonstrated that soybean peroxygen-
ase catalyzes exclusively the heterolytic cleavage of the
hydroperoxide bond yielding 13-hydroxy-9,11,15-octa-
decatrienoic acid (13 HOT), the corresponding alcohol
and an oxoferryl intermediate state of the enzyme.
6
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
A
B
1400
f
a
9R,10S
9S,10R
-
-
2000
9R,10S
9S,10R
0
300
g
h
0
10000
b
0
0
40
650
0
9S,10R
Fig. 5. (A) Separation by radio-chiral-phase
HPLC of the two enantiomers of racemic
methyl 9,10-epoxystearate (a) and of the
epoxides resulting from the incubation of
C18:1 with PXG4 in the presence of
cumene hydroperoxide (b), H2O2 (c), 13
HPOD (d) and 13 HPOT (e). (B) Analysis
of radio-chiral-phase HPLC of the epoxide
resulting from the incubation of PXG4
with the [14C]methyl ester derivative of
oleic acid (g). (f) Analysis of racemic 9,10-
epoxystearate. (h), (i), (j), (k) Analysis of
the enantiomers of the two
9R,10S
- c
3000
0
1700
i
j
0
-
5000
d
0
300
12S,13R
12R,13S
0
monoepoxides resulting from the
e
-
800
0
incubation of C18:2 with PXG4 in the
presence of cumene hydroperoxide: (h)
racemic 9,10-epoxy-12-octadecenoate; (i)
9(R),10(S)-epoxy-12-octadecenoate formed
through PXG4 catalysis; (j) racemic 12,13-
epoxy-9-octadecenoate; (k) 12(R),13(S)-
epoxy-9-octadecenoate catalyzed by
PXG4.
300
k
0
0
0
30
Time (min)
60
0
140
Time (min)
This reactive intermediate is then able to epoxidize the
double bonds of 13 HOT before this reaction product
diffuses out of the active site (caged reaction, intramo-
lecular oxygen transfer mechanism). Alternatively, the
oxoferryl intermediate can also catalyze co-oxidation
which involves an intermolecular oxygen transfer
mechanism; this includes the formation of epoxy alco-
hols deriving from 13 HOT after its diffusion out of
the active site and rebinding (Fig. 6). It is of note that
during fatty acid epoxidation such as oleic acid by the
soybean peroxygenase, epoxy alcohol was formed only
as a minor product [28]. As PXG4 actively catalyzed
co-oxidation reactions such as the epoxidation of
mono-unsaturated fatty acids, we investigated whether
this enzyme was able to catalyze the formation of
epoxy alcohols from fatty acid hydroperoxides via
intramolecular or intermolecular transfers of oxygen.
To study the formation of epoxy alcohol in the
absence of any other oxidizable substrate, we incu-
bated purified PXG4 only in the presence of [1-14C]13
HPOT. Two radioactive compounds could be
extracted from the reaction medium and separated
by RP-HPLC (Fig. 7Aa). The first compound had
the same retention time as the standard 13 HOT
(12.7 min) and both presented identical mass spectra
when analyzed by GC-MS. The second compound had
a higher polarity (elution time 5.4 min). GC-MS anal-
ysis revealed that it consisted of two epoxy alcohol
isomers characterized by their distinct fragmentation
patterns. The mass spectrum of the major peak (96%)
(i.e. the methyl ester trimethylsilyl ether derivative of
15,16-epoxy-13-hydroxy-9,11-octadecenoic acid) was
characterized by a molecular ion at m/z 396 and a
peak base at m/z 73 (trimethylsilyl). Prominent ions
were also seen at m/z (relative intensity) 311 (45%)
(M-CH3CH2CHCOHCH2,), 381 (1.2%) (M-15), 367
(2.1%) (M-CH3CH2), 337 (6.2%) (M-COOCH3), 239
(7%) and 157 (8.5%) (cleavage between C9 and C10)
and 221 (3.9%) (239-O). Mass spectra of the minor
peak (4%) (i.e. the methyl ester trimethylsilyl ether
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
7

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
Intra-molecular transfer
of oxygen
13HPOT
13HOT
b
R
O
R
R
OOH
OH
OH
Epoxy-alcohol
a
c
COOH
COOH
Oleic acid
PGX
Linoleic acid
d
COOH
Linolenic acid
16
15
9
10 12 13
S
SO
R
R
OOH
OH
13HPOT
13HOT
Co-oxidation reactions
Fig. 6. Simplified scheme for the reactions catalyzed by peroxygenase. In an initial step peroxygenase catalyzes the reduction of 13 HPOT
to its corresponding alcohol (13 HOT) (pathway a). This process is followed by oxidation steps (pathways b, c and d). The reaction product
can be further transformed by oxidation into epoxy alcohol according to two pathways in competition: (i) an intramolecular oxygen transfer
mechanism (pathway b) involving a 13 HOT molecule which remained bound to the active site after step a, and (ii) an intermolecular oxygen
transfer mechanism (co-oxidation reaction) involving the rebinding of 13 HOT that has diffused out of the active site (pathway c). S stands
for substrate (see inset for the structures of the fatty acids used in this work), SO for oxidized substrate.
derivative of 9,10-epoxy-13-hydroxy-11,15-octadece-
noic acid) showed a similar molecular ion at m/z 396
and base peaks at 73 (trimethylsilyl). Other ions could
be detected at m/z (relative intensity) 171 (84%)
(CH3CH2CH=CHCH2CH), 380 (0.3%) (M-O) and 365
(0.6%) (M-COOCH3), 282 (20%) (M-O-171-73), 311
(5%), 129 (17.5%) and 103 (14%).
alcohol (about 67% of the reaction products) was
observed when PXG4 acted on 13 HPOT in the
absence of competing substrates (Fig. 7A,B). These
data suggested that the oxygen atom originating from
the hydroperoxide can be transferred to 13 HOT
according to two distinct pathways, i.e. intramolecular
and intermolecular, in competition.
Using [14C]13 HPOD as substrate, instead of [14C]13
HPOT, considerably slowed PXG4 activity (Fig. 7Ab).
The epoxy alcohol represented only 4% of the reaction
products when 13 HPOD was incubated in the pres-
ence of PXG4 (Fig. 7Ab). This contrasts with the
accumulation of 15,16-epoxy alcohol (36% of the reac-
tion products) observed, under the same experimental
conditions, in the epoxidation of 13 HOT (Fig. 7Aa).
The reaction mechanism of the soybean peroxygen-
ase predicts that the formation of epoxy alcohol by
epoxidation of the alcohol (formed after the heterolytic
cleavage of the peroxy O–O bond) could be in compe-
tition with the oxidation of other extraneous substrates
(Fig. 6, pathways c and d). To test this point, increas-
ing amounts of linoleic acid were incubated with
PXG4 in the presence of fixed concentrations of [14C]
13 HPOT. As shown in Fig. 7B, increasing concentra-
tions of the free fatty acid resulted in a decreased
formation of the epoxy alcohol deriving from 13
HOT. However, even at high concentrations, linoleic
acid could not totally abolish the formation of the
epoxy alcohol. Conversely, the highest yield of epoxy
Epoxidation of linoleic acid into diepoxides (peak 7),
12,13-monoepoxide (peak 8) and 9,10-monoepoxide
(peak 9) (Fig. 7Ca) occurred during the formation of
the epoxy alcohol and competed with the epoxidation
of 13 HOT (Fig. 8B). In order to evaluate the influence
of hydroperoxides on the epoxidation rate, we incu-
bated [1-14C]linoleic acid in the presence of PXG4 and
the same concentration (200 lM) of cumene hydroper-
oxide, 13 HPOD and 13 HPOT as co-oxidants. Analy-
sis of the products showed that the formation of the
final product (i.e. the diepoxide) is faster in the pres-
ence of cumene hydroperoxide (Fig. 7Cb) than fatty
acid hydroperoxides such as 13 HPOD (Fig. 7Cc) or
13 HPOT (Fig. 7Ca). This result indicates that hydro-
peroxides that cannot be further oxidized by PXG4
after their reduction into alcohols (e.g. H2O2, cumene
hydroperoxide) are strong co-oxidants for epoxidation
reactions. In addition, such results strongly suggest that
in the presence of other suitable substrates (e.g. in the
context of plant cells) co-oxidation reactions such
as fatty acid epoxidations might be predominantly
catalyzed by PXG4.
8
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
A
C
Fig. 7. (A) Radio-RP-HPLC analysis of the
products resulting from 2 min incubation
of [14C]13 HPOT (a) and [14C]13 HPOD (b)
with PXG4 in the same experimental
conditions. (B) Inhibition of the formation
of epoxy alcohol by increasing amounts of
linoleic acid. (C) Co-oxidation of [1-14C]
linoleic acid by PXG4 in the presence of
(a) 13 HOT, (b) cumene hydroperoxide and
(c) 13 HPOD. The products of the reaction
after incubation for 2 min at 27 °C were
analyzed by radio-RP-HPLC. Peak 1,
B
mostly 15,16-epoxy-13-hydroxy-9,11-
octadecenoic acid; peak 2, 13 HOT; peak
3, 13 HPOT; peak 4, mostly 9,10-epoxy-
13-hydroxy-11-octadecanoic acid; peak 5,
13-hydroxy-9,11,15-octadecatrienoic acid;
peak 6, 13 HPOD; peak 7, 9,10-12,13-
diepoxide; peak 8, 12,13-monoepoxide;
peak 9, 9,10-monoepoxide.
A
B
C
6
4
2
0
0.6
0.6
0.4
0.2
0
0.4
0.2
0
F1
F2
F3
F4
Fig. 8. Quantitative RT-PCR was used to access the expression of PXG4 in RNA samples extracted from various tissues (A); flowers at
different stages (B) F1 tiny flower buds, F2 flower buds just before anthesis, F3 just open flowers, F4 fully open flowers; (C) different
flower parts. The Ct (threshold cycle) values obtained for PXG4 were normalized to the Ct values obtained for four reference genes
(At2g28390, At4g34270, At1g13320, At4g26410) whose stable expression in Arabidopsis tissues is known [58]. Relative expression was
calculated with the specific efficiency (E) of each primer pair by using the EDCt method. The given results represent mean values SE of
three biological replicates.
Expression of PXG4 in plant tissues
Discussion
In contrast with the seed-specific PXG1, class II caleo-
sin gene was found expressed in non-oil storage tissues
[32] and PXG4 particularly in flowers [34]. As well,
quantitative RT-PCR confirmed PXG4 expression
among all the tissues of Arabidopsis, including flowers
(Fig. 8A). This expression was especially important in
the petals of fully open flowers (Fig. 8B,C) and to a
lesser extent in the inflorescence stems.
Most of the genes encoding the oxylipin biosynthetic
enzymes have been identified in plants, with the
notable exception of those responsible for the forma-
tion of EFAs. Our present work indicates that recom-
binant class II caleosin PXG4, expressed in yeast,
catalyzes efficient hydroperoxide-dependent fatty acid
epoxygenation reactions and identifies PXG4 as a new
Arabidopsis peroxygenase-type caleosin [32]. Several
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
9

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
lines of evidence are also in favor of the participa-
tion of PXG4 in fatty acid epoxidation in vivo in
Arabidopsis.
epoxidized by PXG4 (Fig. 4B), leading in the case of
linoleic acid to the formation of the opposite stereoiso-
mer, i.e. 12(R),13(S)-epoxy-9(Z)-octadecenoic acid or
(À)-vernolic acid. This suggests that plants possess
evolutionarily divergent sets of enzymes facilitating the
formation of distinct fatty acid epoxides that could
play pivotal roles in different stages of development.
In this context, it is interesting to note that a peroxy-
genase and a cytochrome P450 were reported to
coexist in Euphorbia lagascae seeds, but only peroxy-
genase activity was detectable in seedlings 2 days after
germination [47].
First, PXG4 is expressed in most organs of A. thali-
ana [36] (Fig. 8). So far, peroxygenases (recently classi-
fied as EC 1.11.2.3) were only found in seeds and
young seedlings, leading to the conclusion that these
enzymes are seed-specific. Thus, peroxygenase activity
has been shown to decrease during germination in soy-
bean and oat [42]. The caleosins identified up to now
to be endowed with peroxygenase activity, i.e. Arabid-
opsis PXG1 and PXG2, oat AsPXG1 and rice EFA27,
are also mostly present in seeds and seedlings
[32,33,43]. Moreover, all our attempts to measure per-
oxygenase activity in green tissues have failed. This
could be due either to a lack of expression of this
enzyme in such tissues or to its inhibition by endoge-
nous inhibitors or/and competing substrates such as
carotenoids [44], released in the plant extracts. The
presence of PXG4 transcripts in stems, roots and flow-
ers is thus a strong indication that peroxygenase exists
and can function in organs other than seeds.
Third, PXG4 differs from all the other peroxygenases
by a stereospecificity that has never been described to
date, giving access to the fatty acid epoxide stereoi-
somers found in planta. From a physiological point of
view, the fact that polyunsaturated fatty acids are
actively epoxidized by PXG4 with a high stereoselec-
tivity indicates that the resulting EFAs could be physi-
ological metabolites. Indeed, PXG4 is able to catalyze
the formation of the various anti-fungal compounds
originating from linoleic and linolenic acids which
have been found in rice, i.e. the five epoxy derivatives
of linoleic and linolenic acids and the four allyl
alcohols originating from the reduction of 9- and
13-hydroperoxides deriving from these two unsatu-
rated fatty acids. Importantly, the rice epoxides have
been determined by Kato et al. [19] to have the (R),
(S) configuration, i.e. precisely the one obtained with
PXG4 (Fig. 5). Altogether these results suggest that
the anti-fungal oxylipins might be produced by a rice
peroxygenase orthologous to PXG4 or to another
member of the class II caleosins. Indeed, since PXG4
shares a high sequence identity with class II caleosins
(81%, 64% and 55% with At1g70670, At1g23240 and
At1g23250 respectively), it is tempting to ascribe simi-
lar catalytic activity to these proteins. However, the
identification of PXG4 as the unique fatty acid epoxy-
genase present in Arabidopsis or as the first character-
ized member of a small family of epoxidases remains
to be clarified. Noteworthy, the allyl alcohols isolated
from rice plants were found to be mostly racemates
[19], indicating that not only the hydroperoxides
produced by lipoxygenases but also those formed by
nonenzymatic peroxidation might be used as co-oxi-
dants by the rice enzyme. Although hydroxy
fatty acids that are formed both enzymatically and
non-enzymatically accumulate in Arabidopsis upon
infection with Pseudomonas syringae [48], the involve-
ment of PXG4 in the defence responses of A. thaliana
to pathogens remains to be determined. Recent gain
and loss of function experiments have suggested that
PXG4 might be involved in drought and salt tolerance
Second, in contrast to oat class I caleosin AsPXG1
for which oleic acid is the preferred lipid substrate
[33], PXG4 shows a high catalytic efficiency with phys-
iological polyunsaturated fatty acids such as linoleic
and linolenic acids, which are oxidized into their corre-
sponding monoepoxides and diepoxides in vitro. Thus,
epoxidation of linoleic acid by recombinant PXG4 in
the presence of cumene hydroperoxide (which yields
preponderantly the diepoxide) proceeds via a very fast
initial formation of the monoepoxide, which is further
oxidized into the diepoxide but at a much reduced rate
(about two orders of magnitude slower than the rate
of monoepoxidation). In vivo, besides being a slower
process, the formation of the diepoxide will also be in
competition with the hydrolysis of the monoepoxide
by EHs. Both parameters could explain why diepox-
ides are rarely detected in planta whereas PXG4 is able
to catalyze their formation in vitro. Strikingly, the
epoxidation activity of PXG4 exhibited regioselectivi-
ties and stereoselectivities that are distinct and comple-
mentary to those of the other known plant epoxidases
which oxidize mainly the C-12,13 double bond of
unsaturated fatty acids. For example, only vernolic
acid accumulates in Arabidopsis seeds expressing
Crepis D12-epoxygenase [11,45]. Similarly, tobacco
callus and somatic soybean embryos expressing the
Euphorbiaceae CYP726A1 [15] and recombinant
Arabidopsis CYP77A4 mainly formed the 12(S),13(R)-
epoxy derivative of linoleic acid, i.e. (+)-vernolic acid
[46]. In sharp contrast, the C-12,13 double bond of
unsaturated fatty acids is the least favored unsaturation
10
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
by acting as a negative regulator of abscisic acid
responses, possibly through the formation of at pres-
ent uncharacterized oxylipins [36]. Actually compre-
hension of the links between the biological roles and
catalytic activities of caleosins remains hampered
mainly by the difficulty in uncovering their physiologi-
cal substrates. Since in this work PXG4 was shown to
oxidize unsaturated fatty acids into EFAs, the study
of the relevance of these metabolites in A. thaliana
stress tolerance will be of importance. Recent studies
have reported that caleosins perform different func-
tions in plant cells. For example, AtPXG1 is involved
in the degradation of oil-body storage lipids during
seed germination [49] while RD20 (At2g33380) and
PXG4 play a role in respectively a positive and nega-
tive regulation of abscisic acid responses [36,50]. The
molecular mechanisms underlying such biological
functions remain unknown, however, but one can
speculate that they might be linked to the enzymatic
activity of caleosins and to their ability to synthesize
oxylipins. The evidence presented here that PXG1 and
PXG4 are distinct oxygenases (Fig. 2), which yield dif-
ferent classes of oxygenated products, sustains this
hypothesis. Accordingly, a functional specialization of
oxylipins depending on their molecular structure has
been suggested recently for the 9-lipoxygenase-derived
oxylipins [51].
Subcellular fractionation of yeast
Yeast fractionation was carried out at 4–5 °C essentially as
previously described [32] with some modifications. Yeast
cells (from 50 mL culture) were collected by centrifugation
at 3000 g for 5 min. The pelleted cells were resuspended in
cold buffer A (100 mM KH2PO4, pH 8, 300 mM NaCl, and
yeast protease inhibitor cocktail (Roche Diagnostics,
Boulogne Billencourt, France) at 1 tabletÁ(50 mL)À1). An
equal volume of glass beads (0.45 mm diameter) was added
and the yeast cells were disrupted by vortexing for 30 s
(15-fold) with a lag time of 30 s in ice between each stir.
Homogenate was removed and the glass beads were washed
three times with buffer A. Homogenate and washings were
centrifuged initially at 3000 g for 5 min, and then at
100 000 g for 1 h. After centrifugation, the three fractions
(i.e. LDs at the top, cytosol in the middle fraction and the
microsomal pellet) were resuspended in the same volume of
buffer, which was used immediately for enzymatic activity
measurement or for PXG4 purification, or analyzed by
western blot as described below.
Western blot analysis
Proteins were separated by SDS gel electrophoresis on 14%
polyacrylamide gels and transferred to Immobilon FL
membranes (Millipore, Molsheim, France) at 20 V for
90 min in
a transfer apparatus (Bio-Rad, Marnes-la-
Coquette, France) using a buffer consisting of 50 mM Tris/
HCl, 50 mM boric acid. The membranes were blocked with
5% skim milk in NaCl/Pi (137 mM NaCl, 27 mM KCl,
4.3 mM Na2HPO4, 1.4 mM KH2PO4, pH 7.5) for 60 min,
rinsed and then incubated overnight at 4 °C with anti-His6
mouse monoclonal IgG (Roche, Boulogne-Billencourt,
France) diluted 1 : 5000 in 5% milk, 0.1% Tween 20 and
NaCl/Pi. After extensive washing in NaCl/Pi containing
0.1% Tween 20, anti-mouse antibody conjugated to peroxi-
dase (Roche) was diluted 1 : 5000 and used as a secondary
antibody. Blots were developed using the ECL kit from
Pierce (Thermo Fisher Scientific, Courtaboeuf, France).
To sum up, until recently the significance of fatty
acid epoxidation in plants was largely impeded by the
lack of genes known to encode efficient epoxidases.
The present identification of such a gene which
encodes a non-seed-specific peroxygenase able to
potently catalyze the epoxidation of physiological
unsaturated fatty acids will now permit the manipula-
tion of EFA levels in plants to assess their biological
functions.
Materials and methods
Enzymatic activity
Functional expression of At1g70670 in yeast
Peroxygenase activity was measured with aniline as the sub-
strate during the purification procedure [32]. The epoxida-
tion of [14C] unsaturated fatty acids was performed as
described in an earlier study [28]. The specificity constants
(or V/Km) were determined under [S]0 << Km experimental
conditions. Under these conditions, the integrated rate
equation for irreversible single-substrate reactions is simpli-
fied into a pseudo-first-order rate equation which yields the
ratio V/Km [55]. For a typical assay, the enzyme was
pre-equilibrated for 1 min at 27 °C in 0.1 M potassium
phosphate buffer (pH 7.4) containing cumene hydroperox-
ide (200 lM) in a final volume of 400 lL. The reaction was
Full-length At1g70670 DNA (pda03699 provided by
RIKEN BRC) [52,53] was PCR amplified using the following
primers: 5′-CGGGATCCATGGCTTCCTCTATTTCCAC
TGG-3′ and 5′-GCTCTAGAGTGGTGGTGGTGGTGG
TGTTAGTTATGGATGTTTCTTAGAAGTTTTAGAAG
ATC-3′. The amplified product was inserted into the
BamH1 and XbaI restriction sites of pVT102U plasmid
[54] in DH5a Escherichia coli and subcloned in yeast
Wa6 after sequencing. Expressions and purifications of
At4g26740 and At1g77670 on a nickel-nitrilotriacetic acid
Superflow column (Qiagen, Courtaboeuf, France) were
performed as previously described [32].
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
11

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
initiated by adding 2 lM of [14C] mono-unsaturated or poly-
unsaturated fatty acids. At given times, 50 lL was with-
drawn and the reaction was stopped with 50 lL of
acetonitrile. The mixture was then applied to a silica gel
TLC plate (60 F254; Merck Millipore, Guyancourt, France)
developed in n-hexane/diethylether/formic acid (50 : 50 : 1).
Radioactivity associated with bands corresponding to the
epoxides and the residual fatty acids was determined. In the
absence of enzymatic activity (boiled enzyme) or of cumene
hydroperoxide, no transformation of the unsaturated fatty
acids was detectable. Under these conditions the reaction
rates were linear with regard to enzyme concentration.
0.3 mLÁminÀ1
)
for methyl cis-9,10-epoxystearate. The
enantiomers of the methylated monoepoxides of [14C]lino-
leic acid were separated on the same column with a solvent
mixture of n-hexane/isopropanol (100 : 0.04 at 0.8 mLÁminÀ1).
In this system, the elution times for 9S,10R-12-octadeceno-
ate, 9R,10S-12-octadecenoate, 12S,13R-epoxy-9-octadece-
noate and 12R,13S-9-octadecenoate were 77.6, 95.6, 100.8
and 115.1 min respectively. The identification of all the
epoxides and epoxy alcohols was performed using larger
incubation volumes and radioactive tracers. Products were
extracted and their identity was confirmed by GC-MS
[28,46,56,57].
Epoxides resulting from the incubation of C18:1 (with
the double bond at a different location in the carbon chain)
were methylated with etheral diazomethane before their
identification by GC-MS analysis (see below). The mass
spectrum of the epoxide resulting from oxidation of oleic
acid was identical (after methylation) to that of authentic
9,10-epoxyoctadecanoate methyl ester, i.e. each spectrum
featured a molecular ion at m/z 312 and prominent ions at
155, base peak (M-157, loss of.(CH2)7COOCH3), 199
(M-133, loss of.(CH2)7CH3) and 281 (M-31, loss of.OCH3).
Capillary gas chromatography analysis of the mono-
unsaturated and polyunsaturated fatty acid methyl esters
and their epoxide derivatives was performed on a Varian
Star chromatography workstation equipped with a DB1
coated fused column (250 lm film thickness, 30 m, 0.32 lm
internal diameter; J.W. Scientific, Agilent Technologies,
Massy, France). Analysis was performed (H2, 2 mLÁminÀ1
)
with the following temperature program: 6 °C ÁminÀ1 from
60 to 120 °C, followed by 5 °CÁminÀ1 from 120 to 240 °C
and by 10 °CÁminÀ1 from 240 to 300 °C. GC analysis was
performed on reaction products resulting from three differ-
ent biological experiences.
Analytical procedures
The GC-MS analysis was performed on an Agilent
5973 N apparatus with an ionizing energy of 70 eV. The
sample was injected directly in splitless mode (injector tem-
perature 250 °C) into a DB-5-coated fused column (30 m,
0.25 mm internal diameter; J.W. Scientific) with a tempera-
ture program of 3 °CÁminÀ1 from 100 to 280 °C, followed
by 10 min at 280 °C (m/z = 50–700).
Radioactivity was measured on TLC plates with a radioiso-
tope thin-layer analyzer Rita Star, and peak integration
was obtained using the Gina-Star TLC software (Raytest,
Straubenhardt, Germany).
HPLC analysis was performed under isocratic conditions
using a Shimadzu instrument coupled with a radiomatic
500 TR analyzer (Packard Instrument Co. PerkinElmer,
Courtaboeuf, France). Peak integration was obtained using
FLO-ONE software. SP-HPLC separation of the monoepox-
ide positional isomers from [14C]linoleic acid was per-
formed at room temperature on a lPorasil (5 lm) column
(3.9 9 300 mm; Waters, Darmstadt, Germany) with a silica
guard cartridge using a solvent mixture of n-hexane/isopro-
panol/acetic acid (995 : 5: 0.5 at 1.5 mLÁminÀ1). In this sys-
tem, the elution times for 12,13-epoxy-9-octadecenoic acid
and 9,10-epoxy-12-octadecenoic acid were 14.20 and
17.80 min respectively. The same chromatographic condi-
tions also enabled 12,13-epoxy-9,15-octadecadienoate,
15,16-epoxy-9,12-octadecadienoate and 9,10-epoxy-12,15-
octadecadienoate to be separated from [14C]linolenic acid
at 14.60, 17.80 and 19.20 min respectively. Diepoxides from
[14C]linolenic acid were separated by reverse phase on a Li-
crospher 100 RP-18 (5 lm) column (8 9 125 mm; Agilent
RNA isolation and quantitative RT-PCR
Fresh material was harvested from at least three different
plants and immediately frozen in liquid nitrogen. Total
RNA was precipitated and extracted by the classical
lithium chloride–phenol protocol, treated with DNAse I
(Fermantas, Thermo Fisher Scientific, Courtaboeuf,
France) according to the manufacturer’s instructions and
further purified with the RNeasy plant mini kit (Qiagen)
according to the manufacturer’s instructions.
cDNA was synthesized with SuperScript III Reverse
Transcriptase (Invitrogen, Life Technologies, Saint Aubin,
France) in the presence of Oligo(dT)23 (Sigma-Aldrich
Chemie S.a.r.l., Saint-Quentin Fallavier, France) from 2 lg
of total RNA. The resulting cDNAs were then 10-fold
diluted. The oligonucleotides used for PXG4 were
At1g70670, 5′-TCAAACAACTCCTAAAGGCAAA-3′, 5′-
CCACTCCGTATACCCTGCAA-3′. The vacuolar protein
MONI (At2g28390, 5′-GGATTTTCAGCTACTCTTCAA
GCTA-3′, 5′-CTGCCTTGACTAAGTTGACACG-3′), the
TIP41-like protein (At4g34270, 5′-GAACTGGCTGACAA
TGGAGTG-3′, 5′-ATCAACTCTCAGCCAAAATCG-3′),
the PP2AA3 (At1g13320, 5′-GACCGGAGCCAACTAGG
Technologies, Interchim, Montluc¸ on, France) using a sol-
vent mixture of methanol/water/acetic acid (75 : 25 : 0.05
at 1 mLÁminÀ1). The resolution of the enantiomers of
14C-labeled methyl epoxy fatty acid derivatives was per-
formed on a Chiralcel OB column (4.6 9 250 mm; Baker
Chemical Co., Interchim, Montluc¸ on, France) with a
solvent mixture of n-hexane/isopropanol (98.5 : 1.5 at
12
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
AC-3′, 5′-AAAACTTGGTAACTTTTCCAGCA-3′) and
the conserved protein UCP022280 (At4g26410, 5′-AGCTGA
AGTGGCTTCCATGA-3′, 5′-GGTCCGACATACCCATGA
TCC-3′) were taken as reference genes to normalize the
expression of PXG4.
8 Krewson CF (1967) Naturally occurring epoxy oils.
J Am Oil Chem Soc 45, 250–256.
9 Budziszewski GJ, Croft KP & Hildebrand DF (1996)
Uses of biotechnology in modifying plant lipids. Lipids
31, 557–569.
10 Sharmin E, Ashraf SM & Ahmad S (2007) Synthesis,
characterization, antibacterial and corrosion properties
of epoxides, epoxy-polyols and epoxy-polyurethane
coating from linseed and Pongamia glabra seed oils.
Int J Biol Macromol 40, 407–422.
Quantitative RT-PCR plates were prepared with a Biomek
3000 (Beckman Coulter, Villepinte, France) and run on a
LightCycler 480 II (Roche). Each reaction was prepared
using 2 lL cDNA, 5 lL of LightCycler 480 SYBR Green I
Master (Roche) and 250 nM of forward and reverse primers
in a total volume of 10 lL. The amplification profile con-
sisted of 95 °C for 10 min and 45 cycles (95 °C denatur-
ation for 10 s, annealing at 60 °C for 15 s, extension at
72 °C for 15 s), followed by a melting curve analysis from
55 to 95 °C to check the specificity of primer pairs. All
reactions were performed in triplicate.
11 Lee M, Lenman M, Banas A, Bafor M, Singh S,
Schweizer M, Nilsson R, Liljenberg C, Dahlqvist A,
Gummeson PO et al. (1998) Identification of non-heme
diiron proteins that catalyze triple bond and epoxy
group formation. Science 280, 915–918.
12 Singh S, Thomaeus S, Lee M, Stymne S & Green A
(2001) Transgenic expression of a D12-epoxygenase
gene in Arabidopsis seeds inhibits accumulation of
linoleic acid. Planta 212, 872–879.
Acknowledgements
The authors are grateful to Francis Schuber (Faculty
of Pharmacy, Strasbourg-Illkirch) for critical reading
of the manuscript and to Anabella Ivancich (CNRS
UMR 8221, CEA-Saclay, France) for helpful
discussions. Many thanks also to Alain Rahier (IBMP-
CNRS-UPR2357, Strasbourg) for help with the
GC-MS apparatus. This work was supported by
FNRS of Belgium (FRFC 2.4.563.08-F) to M.-L.F.
13 Rezzonico E, Moire L, Delessert S & Poirier Y (2004)
Level of accumulation of epoxy fatty acid in
Arabidopsis thaliana expressing a linoleic acid
D12-epoxygenase is influenced by the availability of
the substrate linoleic acid. Theor Appl Genet 109,
1077–1082.
14 Bafor M, Smith MA, Jonsson I, Stobart K & Stymne S
(1993) Biosynthesis of vernolate (cis-12-epoxyoctadeca-
cis-9-enoate) in microsomes preparations from
developing endospems of Euphorbia lagascae. Arch
Biochem Biophys 303, 145–151.
References
15 Cahoon EB, Ripp KG, Hall SE & McGonicle B (2002)
Transgenic production of epoxy fatty acids by
expression of a cytochrome P450 enzyme from
Euphorbia lagascae seed. Plant Physiol 128, 615–624.
16 Kolattukudy PE (1980) Biopolyesters of plants: cutin
and suberin. Science 208, 990–1000.
1 Pfister SL, Gauthier KM & Campbell WB (2010)
Vascular pharmacology of epoxyeicosatrienoic acids.
Adv Pharmacol 60, 27–59.
2 Morisseau C, Inceoglu B, Schmelzer K, Tsai HJ, Jinks
SL, Hegedus CM & Hammock BD (2010) Naturally
occurring monoepoxides of eicosapentaenoic acid and
docosahexaenoic acid are bioactive antihyperalgesic
lipids. J Lipid Res 51, 3481–3490.
17 Schweizer P, Felix G, Buchala A, Muller C & Metraux
JP (1996) Perception of free cutin monomers by plant
cells. Plant J 10, 331–341.
3 Panigrahy D, Edin ML, Lee CR, Huang S, Bielenberg
DR, Butterfield CE, Barnes CM, Mammoto A,
Mammoto T, Luria A et al. (2012) Epoxyeicosanoids
stimulate multiorgan metastasis and tumor dormancy
escape in mice. J Clin Invest 122, 178–191.
18 Schweizer P, Jeanguenat D, Whitacre D, Metraux JP &
Mosinger E (1996) Induction of resistance in barley
against Erysiphe graminis f.sp. hordei by free cutin
monomers. Physiol Mol Plant Pathol 49, 103–120.
19 Kato T, Yamaguchi Y, Namai T & Hirukawa T (1993)
Oxygenated fatty acids with anti-rice blast fungus
activity in rice plants. Biosci Biotechnol Biochem 57,
283–287.
4 Van Rollins MR, Baker RC, Sprecher HW & Murphy
RC (1984) Oxidation of docosahexaenoic acid by rat
liver microsomes. J Biol Chem 259, 5776–5783.
5 Capdevila JH, Falck JR & Harris RC (2000)
Cytochrome P450 and arachidonic bioactivation:
molecular and functional properties of the arachidonate
monoxygenase. J Lipid Res 41, 163–181.
20 Namai T, Kato T, Yamaguchi Y & Hirukawa T (1993)
Anti-rice blast activity and resistance induction of C-18
oxygenated fatty acids. Biosci Biotechnol Biochem 57,
611–613.
6 Zeldin DC (2001) Epoxygenase pathways of arachidonic
acid metabolism. J Biol Chem 276, 36059–36062.
7 Spector AA (2009) Arachidonic acid cytochrome P450
epoxygenase pathway. J Lipid Res 50 (Supplement),
S52–S56.
21 Wagner K, Inceoglu B, Gill SS & Hammock BD (2011)
Epoxygenase fatty acids and soluble epoxide hydrolase
inhibition: novel mediators of pain reduction. Agric
Food Chem 59, 2816–2824.
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
13

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
22 Ingraham RH, Gless RD & Lo HY (2011) Soluble
epoxide hydrolase inhibitors and their potential for
treatment of multiple pathologic conditions. Curr Med
Chem 18, 587–603.
35 Mayfield JA, Fiebig A, Johnstone SE & Preuss D
(2001) Gene families from the Arabidopsis thaliana
pollen coat proteome. Science 292, 2482–2485.
36 Kim YY, Jung KW, Yoo KS, Jeung JU & Shin JS
(2011) A stress-responsive caleosin-like protein,
AtCLO4, acts as a negative regulator of ABA responses
in Arabidopsis. Plant Cell Physiol 52, 874–884.
37 Chen JCF, Tsai CCY & Tzen JTC (1999) Cloning and
secondary structure analysis of caleosin, a unique
calcium-binding protein in oil bodies of plant seeds.
Plant Cell Physiol 40, 1079–1086.
23 Kiyosue T, Beetham JK, Pinot F, Hammock BD,
Yamaguchishinozaki K & Shinozaki K (1994)
Characterization of an Arabidopsis cDNA for a soluble
epoxide hydrolase gene that is inducible by auxin and
water stress. Plant J 6, 259–269.
24 Stapleton A, Beetham JK, Pinot F, Garbarino JE,
Rockhold DR, Friedman M, Hammock BD & Belknap
WR (1994) Cloning and expression of soluble epoxide
hydrolase from potato. Plant J 6, 251–258.
25 Pinot F, Bosch H, Salaun JP, Durst F, Mioskowski C
& Hammock BD (1997) Epoxide hydrolase activities in
the microsomes and the soluble fraction from Vicia
sativa seedlings. Plant Physiol Biochem 35, 103–110.
26 Ohta H, Shida K, Peng YL, Furusawa I, Shishiyama
J, Aibara S & Morita Y (1990) The occurrence of
lipid-hydroperoxide-decomposing activities in rice and
the relationship of such activities to the formation
of antifungal substances. Plant Cell Physiol 31,
1117–1122.
38 Blee E, Summerer S, Flenet M, Rogniaux H, Van
Dorsselaer A & Schuber F (2005) Soybean epoxide
hydrolase. Identification of the catalytic residues and
probing of the reaction mechanism with secondary
kinetic isotope effects. J Biol Chem 280, 6479–6487.
39 Ziegler J, Wasternack C & Hamberg M (1999)
On the specificity of allene oxide cyclase. Lipids 34,
1005–1015.
40 Hamberg M & Fahlstadius P (1992) On the specificity
of a fatty acid epoxygenase in broad bean (Vicia faba
L.). Plant Physiol 99, 987–995.
41 Blee E, Wilcox AL, Marnett LJ & Schuber F (1993)
Mechanism of reaction of fatty acid hydroperoxides with
soybean peroxygenase. J Biol Chem 268, 1708–1715.
42 Blee E & Durst F (1987) Hydroperoxide-dependent
sulfoxidation catalyzed by soybean microsomes. Arch
Biochem Biophys 254, 43–52.
27 Ohta H, Shida K, Peng YL, Furasawa I, Shishiyama J,
Aibara S & Morita Y (1991) A lipoxygenase pathway is
activated in rice after infection with the rice blast
fungus Magnaporthe grisea. Plant Physiol 97,
94–98.
28 Blee E & Schuber F (1990) Efficient epoxidation of
unsaturated fatty acids by a hydroperoxide-dependent
oxygenase. J Biol Chem 265, 12887–12894.
29 Hamberg M & Hamberg G (1990) Hydroperoxide-
dependent epoxidation of unsaturated fatty acids in the
broad bean (Vicia faba L.). Arch Biochem Biophys 283,
409–416.
43 Frandsen GI, Muller-Uri F, Nielsen M, Mundy J &
Skriver K (1996) Novel plant Ca2+- binding protein
expressed in response to abscisic acid and osmotic
stress. J Biol Chem 271, 343–348.
44 Ishimaru A & Yamazaki I (1977) Hydroperoxide-
dependent hydroxylation involving ‘H2O2-reducible
hemoprotein’ in microsomes of pea seeds. J Biol Chem
252, 6118–6124.
30 Blee E & Schuber F (1993) Biosynthesis of cutin
monomers – involvement of a lipoxygenase/
peroxygenase pathway. Plant J 4, 113–123.
31 Lequeu J, Fauconnier ML, Chammai A, Bronner R &
Blee E (2003) Formation of plant cuticule: evidence for
the occurrence of the peroxygenase pathway. Plant J
36, 155–164.
45 Hatanaka T, Shimizu R & Hildebrand D (2004)
Expression of a Stokesia laevis epoxygenase gene.
Phytochemistry 65, 2189–2196.
46 Sauveplane V, Kandel S, Kastner PE, Ehlting J,
Compagnon V, Werck-Reichhart D & Pinot F (2008)
Arabidopsis thaliana CYP77A4 is the first cytochrome
P450 able to catalyze the epoxidation of free fatty acids
in plants. FEBS J 276, 719–735.
32 Hanano A, Burcklen M, Flenet M, Ivancich A,
Louwagie M, Garin J & Blee E (2006) Plant seed
peroxygenase is an original heme-oxygenase with an
EF-hand calcium binding motif. J Biol Chem 281,
33140–33151.
47 Blee E, Stahl U, Schuber F & Stymne S (1993) Regio-
and stereoselectivity of cytochrome P-450 and
peroxygenase- dependent formation of cis-12,13-epoxy-9
(Z)- octadecenoic acid (vernolic acid) in Euphorbia
lagascae. Biochem Biophys Res Commun 197, 778–784.
48 Grun C, Berger S, Matthes D & Mueller MJ (2006)
Early accumulation of non- enzymatically synthesised
oxylipins in Arabidopsis thaliana after infection with
Pseudomonas syringae. Funct Plant Biol 34, 65–71.
49 Poxleitner M, Rogers SW, Samuels AL, Browse J &
Rogers J (2006) A role for caleosin in degradation of
33 Meesapyodsuk D & Qiu X (2011) A peroxygenase
pathway involved in the biosynthesis of epoxy fatty
acids in oat (Avena sativa L.). Plant Physiol 157,
454–463.
34 Naested H, Frandsen GI, Jauh GY, Hernadez-Pinzon I,
Nielsen HB, Murphy DJ, Rogers JC & Mundy J (2000)
Caleosins: Ca2+-binding proteins associated with lipid
bodies. Plant Mol Biol 44, 463–476.
14
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS

´
E. Blee et al.
PXG4 is an efficient fatty acid epoxygenase
oil-body storage lipid during seed germination. Plant J
54 Verner T, Dignar D & Thomas DY (1987) A family of
yeast expression vectors containing the phage f1
intergenic region. Gene 52, 225–233.
47, 917–933.
50 Aubert Y, Vile D, Pervent M, Aldon D, Randy B,
Simonneau T, Vavasseur A & Galaud JP (2010) RD20,
a stress-inducible caleosin, participates in stomatal
control, transpiration and drought tolerance in
Arabidopsis thaliana. Plant Cell Physiol 51, 1975–1987.
51 Lopez MA, Vicente J, Kulasekaran S, Vellosillo T,
Martinez M, Irigoyen ML, Cascon T, Bannenberg G,
Hamberg M & Castresana C (2011) Antagonistic role
of 9- lipoxygenase-derived oxylipins and ethylene in the
control of oxidative stress, lipid peroxidation and plant
defence. Plant J 67, 447–458.
55 Orsi BA & Tipton KF (1979) Kinetic analysis of
progress curves. Methods Enzymol 63, 159–183.
56 Blee E & Schuber F (1992) Enantioselectivity of the
hydrolysis of linoleic acid monoepoxides catalyzed by
soybean fatty acid epoxide hydrolase. Biochem Biophys
Res Commun 187, 171–177.
57 Orellana-Coca C, Adlercreutz D, Andersson MM,
Mattiasson B & Hatti-Kaul R (2005) Analysis of fatty
acid epoxidation by high performance liquid
chromatography with evaporative light scattering
detection and mass spectrometry. Chem Phys Lipids
135, 189–199.
52 Seki M, Carninc P, Nishiyama Y, Hayashizaki Y &
Shinozaki K (1998) High- efficiency cloning of
Arabidopsis full-length cDNA by biotinylated CAP
trapper. Plant J 15, 707–720.
58 Czechowski T, Stitt M, Altmann T, Udvardi MK &
Scheible WR (2005) Genome-wide identification
and testing of superior reference genes for
transcript normalization in Arabidopsis. Plant Physiol
139, 5–17.
53 Seki M, Narusaka M, Kamiya A, Ishida J, Satou M,
Sakurai T, Nakajima M, Enju A, Akiyama K, Oono Y
et al. (2002) Functional annotation of a full-length
Arabidopsis cDNA collection. Science 296, 141–145.
FEBS Journal (2012) ª 2012 The Authors Journal compilation ª 2012 FEBS
15