LC-MS/MS analysis of epoxyalcohols and epoxides of arachidonic acid and their oxygenation by recombinant CYP4F8 and CYP4F22

Article abstract of DOI:10.1016/j.abb.2009.11.013

CYP4F22 and CYP4F8 are expressed in epidermis, and mutations of CYP4F22 are associated with lamellar ichthyosis. Epoxyalcohols (HEETs) and epoxides (EETs) of 20:4n-6 appear to be important for the water permeability barrier of skin. Our aim was to study the MS/MS spectra and fragmentation of these compounds and to determine whether they were oxidized by CYP4F22 or CYP4F8 expressed in yeast. HEETs were prepared from 15-hydroperoxyeicosatetraenoic acid (15-HPETE), 12-HPETE, and their [²H₈]labeled isotopomers, and separated by normal phase-HPLC with MS/MS analysis. CYP4F22 oxygenated 20:4n-6 at C-18, whereas metabolites of HEETs could not be identified. CYP4F8 formed ω3 hydroxy metabolites of HEETs derived from 12R-HPETE with 11,12-epoxy-10-hydroxy configuration, but not HEETs derived from 15S-HPETE. 8,9-EET and 11,12-EET were also subject to ω3 hydroxylation by CYP4F8. We conclude that CYP4F8 and CYP4F22 oxidize 20:4n-6 and that CYP4F8 selectively oxidizes 8,9-EET, 11,12-EET, and 10,11R,12R-HEET at the ω3 position.

Full text of DOI:10.1016/j.abb.2009.11.013

Archives of Biochemistry and Biophysics 494 (2010) 64–71
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
LC–MS/MS analysis of epoxyalcohols and epoxides of arachidonic acid
and their oxygenation by recombinant CYP4F8 and CYP4F22
T. Nilsson a, I.V. Ivanov b, E.H. Oliw a,
*
a Division of Biochemical Pharmacology, Department of Pharmaceutical Biosciences, Biomedicum, Uppsala University, SE-75124 Uppsala, Sweden
b Institute of Biochemistry, Charité-Universitätsmedizin Berlin, 10117 Berlin, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 October 2009
and in revised form 10 November 2009
Available online 15 November 2009
CYP4F22 and CYP4F8 are expressed in epidermis, and mutations of CYP4F22 are associated with lamellar
ichthyosis. Epoxyalcohols (HEETs) and epoxides (EETs) of 20:4nꢀ6 appear to be important for the water
permeability barrier of skin. Our aim was to study the MS/MS spectra and fragmentation of these com-
pounds and to determine whether they were oxidized by CYP4F22 or CYP4F8 expressed in yeast. HEETs
were prepared from 15-hydroperoxyeicosatetraenoic acid (15-HPETE), 12-HPETE, and their [2H8]labeled
isotopomers, and separated by normal phase-HPLC with MS/MS analysis. CYP4F22 oxygenated 20:4nꢀ6
Keywords:
Epoxyeicosatrienoic acid
Electrospray ionization
Hepoxilins
Ichthyosis
Ion trap mass-spectrometry
Lipid signaling
at C-18, whereas metabolites of HEETs could not be identified. CYP4F8 formed
HEETs derived from 12R-HPETE with 11,12-epoxy-10-hydroxy configuration, but not HEETs derived from
15S-HPETE. 8,9-EET and 11,12-EET were also subject to 3 hydroxylation by CYP4F8. We conclude that
CYP4F8 and CYP4F22 oxidize 20:4nꢀ6 and that CYP4F8 selectively oxidizes 8,9-EET, 11,12-EET, and
10,11R,12R-HEET at the 3 position.
x3 hydroxy metabolites of
x
x
Ó 2010 Published by Elsevier Inc.
Cytochromes P450 (CYPs)1 are of paramount importance for
metabolism of endogenous and exogenous compounds. In humans,
57 CYP enzymes have been identified so far [1], and grouped into fam-
ilies and subfamilies based on amino acid similarities. Members of the
CYP4 family oxidize fatty acids [2], leukotrienes [3,4], and prostaglan-
dins [5], but the physiological function has not been established for
several of them (e.g., CYP4F22, CYP4X1, and CYP4Z1 [6]). It seems
plausible that some of these orphan CYPs may be involved in metab-
olism of lipids formed from arachidonic acid (20:4nꢀ6) by any of three
oxidative systems: cyclooxygenase (COX), lipoxygenase (LOX), and
CYP. The presence of an oxidative pathway and expression of a specific
orphan CYP in a tissue could generate a hypothesis on the catalytic
function of this CYP isoform.
LOX has lately attracted particular medical interest. 12R-LOX is
only expressed in epidermis and in tonsils [7,10], and is up regu-
lated in psoriatic lesions [11]. It transforms 20:4nꢀ6 to 12R-hydro-
peroxyeicosatetraenoic acid (12R-HPETE), which is important for
the development of the water permeability barrier [12–14]. Re-
cently, mutations in the gene for 12R-LOX were linked to a rare
skin disease, lamellar ichthyosis [15,16]. 12R-HPETE is further ste-
reospecifically transformed to an epoxyalcohol, 11(R),12(R)-trans-
epoxy-8(R)-hydroxyeicosa-5(Z),9(E),14(Z)-trienoic acid (8R,11R,-
12R-HEET) (Fig. 1) by the enzyme epidermal LOX-3 (eLOX3) [17].
Mutational studies have also associated eLOX3 with lamellar ich-
thyosis [15,16,18–20], and recently the gene for CYP4F22 was
linked to this disease [9]. Lefévre et al. have proposed that CYP4F22
takes part in bio activation of products derived from 12R-HPETE
and 8R,11R,12R-HEET [9,21], but this hypothesis has not yet been
investigated likely reflecting difficulties in expression of recombi-
nant CYP4F22.
Human epidermis contains 15S-LOX type 1, 12S-LOX, and 12R-
LOX [7]. Skin also contains CYPs involved in the transformation of
vitamin A and vitamin D, and some members of the CYP4 family
with unknown epidermal function [8,9]. Of these enzymes, 12R-
Epidermal 12S-HPETE is formed by platelet-type 12S-LOX in hu-
mans. HEETs derived from 12S-HPETE are found in many tissues
and hepoxilin A3 may influence the secretion of insulin and the re-
lease of serotonin [22,23], but their function in skin is unknown.
20:4nꢀ6 can be oxidized by epoxygenases of the CYP1-3 subfamilies
to four regio isomeric cis epoxides (5,6-epoxyeicosa-8(Z),11(Z),14(Z)-
trienoic acid (5,6-EET), 8,9-epoxyeicosa-5(Z),11(Z),14(Z)-trienoic acid
(8,9-EET), 11,12-epoxyeicosa-5(Z),8(Z),14(Z)-trienoic acid (11,12-EET),
and 14,15-epoxyeicosa-5(Z),8(Z),11(Z)-trienoic acid (14,15-EET)[24].
EETs affect the water permeability barrier in mouse [25], and both
* Corresponding author. Address: Division of Biochemical Pharmacology, Depart-
ment of Pharmaceutical Biosciences, Biomedicum, Uppsala University, P.O. Box 591,
SE-75124 Uppsala, Sweden. Fax. +46 18 4714748.
E-mail address: Ernst.Oliw@farmbio.uu.se (E.H. Oliw).
Abbreviations used: CP, chiral phase; CYP, cytochrome P450; EET, epoxyeicosatri-
1
enoic acid; HEET, epoxy-hydroxyeicosatrienoic acid; eLOX, epidermal lipoxygenase;
HETE, hydroxyeicosatetraenoic acid; HPETE, hydroperoxyeicosatetraenoic acid; HPLC,
high performance liquid chromatography; LOX, lipoxygenase; LC–MS/MS, liquid
chromatography–tandem mass spectrometry; NP, normal phase; PCR, polymerase
chain reaction; PGH, prostaglandin H; RP, reverse phase; TIC, total ion current.
0003-9861/$ - see front matter Ó 2010 Published by Elsevier Inc.
doi:10.1016/j.abb.2009.11.013

T. Nilsson et al. / Archives of Biochemistry and Biophysics 494 (2010) 64–71
65
and Escherichia coli strain (Top10) were from Invitrogen. The en-
hanced avian RT-PCR kit and NADPH were from Sigma. Restriction
enzymes were from New England BioLabs and Fermentas. QIA-
quick gel extraction kits were from Qiagen. Yeast extract and pep-
tone from soybean were from Merck. Taq DNA polymerase was
from Promega. PCR primers were from Cybergene (Huddinge, Swe-
den) and from TIB Molbiol (Berlin, Germany). Sequencing was per-
formed at Uppsala Genomic Center (Rudbeck Laboratories, Uppsala
University). Human cornea was obtained from Uppsala Academic
Hospital (Uppsala Human Ethics Committee; Ups 03-546). Other
solvents were from Merck (Darmstadt, Germany), and were of
pro analysis quality.
20:4n-6
-OOC
-OOC
12R-LOX
12R-HPETE
OOH
eLOX3
Expression of CYP4F8 and CYP4F22
8R,11R,12R-HEET
O
Recombinant CYP4F8 has previously been expressed in Sac-
charomyces cerevisiae strain W(R) [5]. CYP4F22 was now expressed
using the same method. The full length coding region was ampli-
fied by PCR, along with two linker sequences and three additional
adenine residues before the ATG to increase the efficiency. The
amplicon was sequenced, and then subcloned into the yeast
expression vector pYeDP60, between the linker sequences, and un-
der a galactose promoter. The vector was transformed into the S.
cerevisiae strain W(R), with the lithium acetate method. Uracil
and adenine were used for selection. To verify successful transfor-
mation, genomic DNA was isolated [32] and analyzed by PCR (con-
ditions: 10 pmol of each primer, 0.2 mM dNTPs, 3 mM MgCl2, 1.5 U
-OOC
OH
Fig. 1. 20:4nꢀ6 is metabolized to 12R-HPETE by 12R-LOX, and isomerized to
8R,11R,12R-HEET by eLOX3.
x
-hydroxy-EETs and 8R,11R,12R-HEET activate PPAR
important regulatory element of inflammation.
a [26,27], an
CYP4F8 and CYP4F22 are both highly expressed in epidermis,
and in seminal vesicles and testis, respectively [5,8,9,28]. CYP4F8
is known to catalyze 19-hydroxylation of PGH1 and PGH2 formed
by COX-2 of seminal vesicles, a key step in biosynthesis of the
two main prostaglandins of human seminal fluid, 19-hydroxy-
PGE1 and -PGE2 [5]. Since 19-hydroxy-PGs are not found in epider-
mis, an alternative function for CYP4F8 in skin is likely. Recombi-
nant CYP4F8 oxidizes 20:4nꢀ6 and other long chain fatty acids
[5,29], and could be involved in the oxygenation of epoxyalcohols
in the skin.
In an attempt to find the function of CYP4F8 and CYP4F22 in
epidermis, we investigated whether HEETs, EETs, and 20:4nꢀ6
were oxidized by recombinant CYP4F22 and CYP4F8. We also per-
formed a systematic study of the MS/MS spectra and the fragmen-
tation mechanism of HEETs derived from 15- and 12-HPETE.
Taq DNA polymerase in 25 ll 10 mM Tris–HCl (pH 9.0), 50 mM KCl
and 0.1% Triton X-100; 30 reaction cycles, annealing temperature
60 °C). A 325-bp fragment of the CYP4F22 gene was amplified
using 50-CACCACTACCTCGACTTC as forward primer, and 50-GGC
ATTTCTCCTGGTATTC as reverse primer. Yeast was initially grown
on minimal media (casamino acids 1 g/L, yeast nitrogen base 7 g/
L, glucose 20 g/L) for 24 h, before changing to complete media
(yeast extract 10 g/L, bactopeptone 10 g/L, ethanol 16 g/L, glucose
5 g/L). For induction of CYP4F22 and P450 reductase, the culture
was grown on galactose (2% w/v) for 15 h before harvesting the
cells. Microsomes were obtained by breaking the cell walls by glass
beads, differential centrifugation (20,000g for 10 min, +4 °C;
100,000g for 60 min, +4 °C), resuspension of the pellet in a buffer
with 0.05 M Tris–HCl, 20% glycerol, 1 mM EDTA, and homogeniza-
tion. Microsomes were stored at ꢀ80 °C in the same buffer. As neg-
ative controls in PCR and expression studies, we used yeast
transfected with vector without insert and untransfected yeast
(the selection step was excluded).
Materials and methods
Materials
20:4nꢀ6 (>99%) was from Sigma. 12(S)-Hydroperoxyeicosa-
5(Z),8(Z),10(E),14(Z)-tetraenoic acid (12S-HPETE) was from Laro-
dan Fine Chemicals (Malmö, Sweden). 14(S),15(S)-trans-epoxy-
13(S)-hydroxy-5(Z),8(Z),11(Z)-eicosatrienoate (13S,14S,15S-HEET)
methyl ester, 14(S),15(S)-trans-epoxy-11(S)-hydroxy-5(Z),8(Z),12-
(E)-eicosatrienoate (11S,14S,15S-HEET) methyl ester, and 14(S),-
15(S)-trans-epoxy-11(R)-hydroxy-5(Z),8(Z),12(E)-eicosatrienoate
(11R,14S,15S-HEET) methyl ester were from Lipidox (Stockholm,
Sweden). 15S-HPETE and [5,6,8,9,11,12,14,15-2H8]15S-HPETE were
prepared using soybean LOX-1 from 20:4nꢀ6 and [5,6,8,9,11,-
12,14,15-2H8]20:4nꢀ6 (99%), a gift of the late Dr. van Dorp. HPETE
were obtained by vitamin E controlled autoxidation of 20:4nꢀ6
[30]. Rac 16-HETE and 16R-HETE were obtained as described
[31]. The Kromasil-100Si-column was from Dalco ChromTech (Sol-
lentuna, Sweden). The ReproSil Chiral-NR-R, ReproSil Chiral-AM,
and ReproSil 100Si-columns were from Dr. Maisch GmbH (Ammer-
buch, Germany). Octadecasilane silica columns were purchased
from Skandinaviska GeneTec AB (Västra Frölunda, Sweden). Re-
combinant CYP4F8 was expressed in a yeast system, and micro-
somes were prepared as described [5]. Platelets were from
Uppsala Academic Hospital (Uppsala, Sweden). TA cloning kits
Synthesis of 12R-HPETE
Hydroperoxides from 20:4nꢀ6 (100 mg) were produced by
autoxidation [30] under oxygen (36 h, 37 °C) in the presence of
5–10% vitamin E. The reaction was followed by spectrophotometry
at 235 nm. The sample was dissolved in 7% diethyl ether in hexane
and applied on an open column with 6 g of silicic acid (Mallinck-
rodt; activated at 110 °C for >2 h). The column was washed with
100 mL of 7% diethyl ether in hexane and the products were eluted
with 100 mL of 20% diethyl ether in hexane in 10 mL fractions. The
fractions were analyzed by TLC (Kieselgel 60; hexane/ethyl acetate,
50/50), and those containing oxygenated products were combined,
evaporated to dryness, and dissolved in isopropanol/hexane for
purification of 12R-HPETE by preparative CP-HPLC (see below).
Synthesis of HEETs and EETs
HEETs were prepared by treatment of 15S-HPETE, 12S-HPETE
and 12R-HPETE with 1–10 eq. hematin (hemin dissolved in 0.01 M
NaOH) in 0.2 mL 0.1 M K2HPO4 [33,34]. After 10 min the mixture

66
T. Nilsson et al. / Archives of Biochemistry and Biophysics 494 (2010) 64–71
was diluted in water, extracted on a SepPak C18 column, eluted in
ethyl acetate, evaporated to dryness, and dissolved in ethanol. For
preparation of 14(S),15(S)-epoxy-13(R)-hydroxy-5(Z),8(Z),11(Z)-
eicosatrienoic acid (13R,14S,15S-HEET), hematin was added directly
after the synthesis of 15S-HPETE with sLOX-1, the products were ex-
tracted as above, and 13R,14S,15S-HEET was isolated by NP-HPLC.
HEETs of 12S-HPETE and [2H8]12S-HPETE were prepared in situ
using outdated human platelets. 20:4nꢀ6 and [2H8]20:4nꢀ6 were
incubated with the cytosolic fraction of human platelets (1.5 mL)
in the presence of hematin (8 eq., 10 min, 37 °C) [35]. After extrac-
tive isolation, the HEETs were separated and analyzed by NP-HPLC.
the mass spectrometer and 95% to the fraction collector. Mass
spectra of all EETs were in accordance with previously published
data [38].
Metabolites of EETs and 20:4nꢀ6 were analyzed by RP-HPLC as
described [39]. The column (Phenomenex, 150 ꢁ 2 mm) was eluted
at 0.3 mL/min with methanol/water/acetic acid (75/25/0.01). Steric
analysis of 16-HETE was performed on amylose tris(3,5-dimethyl-
phenylcarbamate) coated on silica gel (ReproSil Chiral-AM, 250 ꢁ
2 mm, 5 lm), eluted at 0.15 mL/min with hexane/ethanol/acetic
acid, 95/5/0.1, as described [31]. The effluent from the chiral col-
umns was combined with isopropanol/water, as above, for MS/
MS analysis.
HEETs were typically generated from 50–60
threo 13,14,15-HEET from 60 g 15S-HPETE. Methyl ester deriva-
tives of authentic HEETs were hydrolyzed by treatment with
200 l rat blood plasma (20 min, 37 °C), followed by extractive iso-
lg of 12R-HPETE and
l
Effluent from the columns was subjected to electrospray ioniza-
tion (spray voltage 4.3 kV) in an ion trap mass spectrometer (LTQ,
ThermoFisher) with monitoring of negative ions. Temperature of
the heated transfer capillary was set to 310 °C, and collision energy
to 35% (arbitrary units). Activation time was 30 ms, and the q-value
0.25. Ion isolation width was 1.5 Da with the exception of MS/MS/
MS analysis of HPETE (m/z 335 ? 317 ? full scan), where an isola-
tion width of 5 Da was used in the first selection (m/z 335), and
1.5 Da in the second selection (m/z 317) [37].
l
lation. The HEETs were quantified by MS/MS (m/z 335 ? full scan)
by comparison with known concentrations of authentic HEETs.
Hydrolysis to trihydroxy compounds was not detected by MS/MS
analysis (m/z 353 ? full scan) under our experimental conditions.
EETs were prepared by oxidation of 20:4nꢀ6 with 1.1 eq. m-
chloroperoxybenzoic acid as described [36]. The EETs were sepa-
rated on HPLC (see below) into 14,15-, 11,12-, and 8,9-EET.
Results
Enzyme assay
LC–MS/MS analysis of HEETs
The substrates (usually 10–50
microsomes in 0.1 M KHPO4 buffer (pH 7.4) in a total volume of
100 l (CYPF8), or 250 l (CYP4F22), in the presence of NADPH
lM) were incubated (37 °C) with
HEETs from 15S-HPETE
l
l
HEETs were generated by hematin treatment of 15S-HPETE.
Fig. 2 shows the separation of the HEETs on NP-HPLC (3% isopropa-
nol in hexane). Four HEETs were generated as main products: the
erythro isomer 13S,14S,15S-HEET (peak I) and the threo isomer
13R,14S,15S-HEET (peak II), 11S,14S,15S-HEET (peak III), and
11R,14S,15S-HEET (peak IV). The minor peaks likely consisted of
the HEETs with cis epoxy conformation as judged from their MS/
MS spectra.
(1 mM). As negative controls, reactions without NADPH were used.
The reactions were terminated after 30 min with 1 mL ethanol,
centrifuged (15,000g, 10 min, +4 °C), diluted to 10 mL with water,
and the products were extracted on a SepPak/C18 column, evapo-
rated to dryness, and dissolved in ethanol for RP-HPLC analysis,
or in 3–5% isopropanol in hexane for NP-HPLC analysis (see below).
Analysis of HEETs in cornea
11,14S,15S-HEET and 13,14S,15S-HEET fragmented adjacent to
the hydroxyl group and in the epoxy ring (see Fig. 3A and C). The
fragmentation patterns were confirmed by MS/MS analysis (m/z
335 ? full scan) of the [2H]labeled isotopomers (Fig. 3B and D).
Shared fragments were m/z 317 (Aꢀ – H2O) (in the [2H]labeled iso-
Human cornea was homogenized in KHPO4 buffer (pH 7.4), and
incubated with 20:4nꢀ6 (100
lM, 30 min, 37 °C). The reaction was
stopped with ethanol (4 vols.), centrifuged (5 min, 15,000g, +4 °C),
and the supernatant was evaporated and extracted on a SepPak/C18
column. The extracts were dissolved in isopropanol/hexane for
analysis by NP-HPLC (see below).
I
100
LC–MS/MS analysis
NP-HPLC of HEETs and their corresponding metabolites was per-
80
formed on an analytical silica column (Kromasil, 250 ꢁ 2 mm, 5
l
m,
II
100 Å) using the solvent system hexane/isopropanol/acetic acid, 97/
3/0.01 or 95/5/0.01, and a flow rate of 0.3 mL/min. Post column the
effluent was combined in a T-junction with isopropanol/water (3/
2; 0.2–0.3 mL/min) from a Surveyor MS pump. Preparative purifica-
tion of enantiomers of 13,14S,15S-HEET was performed in the same
III
IV
60
40
20
0
way on a ReproSil 100 Si-column (250 ꢁ 10 mm, 5
lm), and a flow
rate of 2 mL/min; 1 min fractions were collected.
Separation of 12R-HPETE from other HPETEs was achieved on a
m) eluted at 0.6 mL/
ReproSil Chiral-NR-R column (250 ꢁ 4 mm, 8
l
min with 1% isopropanol in hexane as mobile phase [37]. The efflu-
ent was collected in 1 min fractions, which were analyzed by MS/
MS (direct injection). The fractions with 12R-HPETE were evapo-
rated to dryness, dissolved in ethanol, and stored at ꢀ20 °C.
8,9-EET was purified by NP-HPLC on a ReproSil 100 Si-column
10
14
18
22
26
30
Retention Time (min)
Fig. 2. NP-HPLC analysis of HEETs formed from 15S-HPETE by hematin. The four
major products were (I) 13(S)-hydroxy-trans-14,15-epoxyeicosatrienoic acid, (II)
13(R)-hydroxy-trans-14,15-epoxyeicosatrienoic acid, (III) 11(S)-hydroxy-trans-
14,15-epoxyeicosatrienoic acid, (IV) 11(R)-hydroxy-trans-14,15-epoxyeicosatrie-
noic acid. The products were eluted with hexane/isopropanol/acetic acid, 95/5/
0.01, and the effluent was mixed with isopropanol/water, 60/40, and subject to
electrospray ionization.
(250 ꢁ 10 mm, 5
lm), whereas 11,12-EET and 14,15-EET were sep-
arated by preparative RP-HPLC (Phenomenex, 200 ꢁ 8 mm; eluted
at 1.2 mL/min with methanol/water/acetic acid, 80/20/0.1). The
column was connected to a splitter, allowing 5% of the effluent into

T. Nilsson et al. / Archives of Biochemistry and Biophysics 494 (2010) 64–71
67
193
317.3
OH
100
80
60
40
20
0
A
B
324.4
OH
*
100
80
60
40
20
0
+
H
O
-
O
OOC
+
H
-
OOC
*
*
*
*
*
*
*
223
235
205
O
-
OOC
OH
199.3
179.2
235.2
255.3
273.3
173.3
242.3
223.3
193.2
262.4
299.3
305.3
167.3
161.3
131.2
129.3
140
335.2
340
340
300
100
180
220
260
100
180
220
m/z
260
300
140
m/z
167
D
C
+H
O
O
-
OOC
-
OOC
*
*
*
*
*
*
*
*
HO
HO
235
235.2
242.3
100
100
80
60
40
20
0
80
60
40
317.3
324.3
171.2
167.2
191.3
20
0
217.2
198.3
224.3
273.3
280.4
299.3
300
149.2
140
300
180
220
m/z
260
340
100
140
100
340
180
220
m/z
260
Fig. 3. MS/MS analysis of HEETs derived from 15S-HPETE. (A) 13-Hydroxy-trans-14,15-epoxyeicosatrienoic acid (m/z 335 ? full scan) with evidence of keto-enol
tautomerism (see insert). (B) [2H8]13S-Hydroxy-trans-14,15-epoxyeicosatrienoic acid (m/z 343 ? full scan). (C) 11-Hydroxy-trans-14,15-epoxyeicosatrienoic acid. (D)
[2H8]11-Hydroxy-trans-14,15-epoxyeicosatrienoic acid. Deuterium is marked with an asterisk. The inserts show the proposed fragmentations.
topomers m/z2H 324), m/z 299 (317 – H2O) (m/z2H 305–306), m/z
273 (Aꢀ – H2O – CO2) (m/z2H 280), m/z 255 (273 – H2O) (m/z2H
100
262), m/z 235 (Aꢀ – O@CH–(CH2)4–CH3) (m/z2H 242), m/z 217
I
(235 – H2O) (m/z2H 223–224), m/z 191 (235 – CO2) (m/z2H 198),
m/z 173 (191 – H2O) (m/z2H 179–180). The discriminating ions
were m/z 167 (ꢀOOC–(CH2)3–CH@CH–CH2–CH@CH–CH3) (m/z2H
171) for 11,14S,15S-HEET, and m/z 193 (ꢀOC–(CH2)3–CH@CH–
CH2–CH@CH–CH2–CH@CH2) (m/z2H 199) for 13,14S,15S-HEET. In
the spectrum of 13-hydroxy derivative the signal at m/z 205 was
identified as originating from a Payne rearrangement [40] of the
epoxy group during the MS analysis (insert Fig. 3A).
II
III
80
60
40
20
0
IV
HEETs from 12-HPETE
Fig. 4 shows the separation of the HEET products of 12S-HPETE
on NP-HPLC. The four main products were identified as 11,12-
trans-epoxy-10(R/S)-hydroxyeicosatrienoic acid (10,11,12-HEET;
peak I and II), and 11,12-trans-epoxy-8(R/S)-hydroxyeicosatrienoic
acid (8,11,12-HEET; peaks III and IV).
Both mass spectra (m/z 335 ? full scan) (Fig. 5A and C) showed
signals at m/z 317 (Aꢀ – H2O) (m/z2H 324), m/z 299 (317 – H2O) (m/
z2H 305–306), m/z 273 (Aꢀ – H2O – CO2) (m/z2H 280), m/z 255 (273
– H2O) (m/z2H 262), and m/z 195 (ꢀOOC–(CH2)3–CH@CH–CH2–
0
4
8
12
16
Retention Time (min)
Fig. 4. NP-HPLC separation of HEETs formed from 12S-HPETE by hematin. The four
major peaks contained (I) 10(S)-hydroxy-trans-11,12-epoxyeicosatrienoic acid, (II)
10(R)-hydroxy-trans-11,12-epoxyeicosatrienoic acid, (III/IV) 8(S/R)-hydroxy-trans-
11,12-epoxyeicosatrienoic acid. The products were eluted with hexane/isopropa-
nol/acetic acid, 95/3/0.01, and the effluent was mixed with isopropanol/water and
subject to electrospray ionization.

68
T. Nilsson et al. / Archives of Biochemistry and Biophysics 494 (2010) 64–71
CO 2
167
B
127
+H
211
O
O
A
-
-
*
*
OOC
OOC
*
*
*
*
*
*
OH
OH
324.3
273.3
100
80
60
40
20
0
100
80
60
40
20
0
195
317.3
280.3
167.2
195.2
255.3
172.3
211.2
216.3 262.3
299.3
300
127.3
140
200.3
180
306.4
129.2
100
180
220
m/z
260
340
220
m/z
260
300
340
100
140
317.2
C
D
100
80
60
40
20
0
153
OH
O
+H
+H
OH
324.3
100
80
60
40
20
0
O
-
OOC
-
*
*
OOC
*
*
*
*
*
*
195
183
296.3
200.2
291.3
300
195.2
273.3
280.3
187.2
255.3
183.3
180
262.3
153.3
140
157.3
140
100
220
m/z
260
340
260
300
180
220
340
100
m/z
Fig. 5. MS/MS analysis of HEETs derived from 12S-HPETE. (A) 8-Hydroxy-trans-11,12-epoxyeicosatrienoic acid (m/z 335 ? full scan). (B) [2H8]8-Hydroxy-trans-11,12-
epoxyeicosatrienoic acid (m/z 343 ? full scan). (C) 10-Hydroxy-trans-11,12-epoxyeicosatrienoic acid. (D) [2H8]10-Hydroxy-trans-11,12-epoxyeicosatrienoic acid. Deuterium
is marked with an asterisk. The inserts show the proposed fragmentations.
C(OH)@CH–CH@CH2 and ꢀOOC–(CH2)3–CH@CH–CH2–CH@CH–
*
100
80
60
40
20
0
CH(OH)@CH2, respectively) (m/z2H 200). 8-Hydroxy-11,12-epoxye-
icosatrienoic acid also showed characteristic signals at m/z 127
(ꢀOOC–(CH2)3–CH@CH–CH3) (m/z2H 129), and m/z 211 (ꢀOOC–
(CH2)3–CH@CH–CH2–CH@CH–CH(OH)–CHO) (m/z2H 216). 10-Hy-
droxy-11,12-epoxyeicosatrienoic acid shows signals at m/z 153
(ꢀOC–(CH2)3–CH@CH–CH2–CH@CH2) (m/z2H 156–157), and m/z
183 (ꢀOOC–(CH2)3–CH@CH–CH2–CH@CH–CH2–OH) (m/z2H 187).
As expected, the MS/MS spectra of HEETs derived from 12R-HPETE
and 12S-HPETE were virtually identical.
II
I
III
Formation of HEETs from 20:4nꢀ6 by human cornea
3
4
5
6
7
Several products were detected by NP-HPLC–MS/MS (m/z
335 ? full scan) from human cornea when incubated with
20:4nꢀ6 (Fig. 6). Peak I was identified as 13S,14S,15S-HEET and
peak II as 13R,14S,15S-HEET, as judged from the MS/MS spectra
and retention times above. Peak III likely contained the two stereo-
isomers of 11,14S,15S-HEET, which were only partly resolved.
Retention Time (min)
Fig. 6. NP-HPLC analysis of HEETs formed from 20:4nꢀ6 by homogenates of human
cornea. Materials in peaks I and II were identified as isomers of 13,14,15-HEET, and
peak III as 11,14,15-HEET by MS/MS analysis (m/z 335 ? full scan). Peak denoted
with an asterisk labels unrelated compounds. The products were eluted with
hexane/isopropanol/acetic acid, 95/5/0.01, and the effluent was mixed with
isopropanol/water, 60/40, and subject to electrospray ionization.
Oxidation of EETs and HEETs by CYP4F8
Oxidation of EETs
CHO) in both mass spectra (m/z 335 ? full scan). We could not de-
tect any dihydroxy products, which validates that the epoxides
were not subjected to epoxide hydrolase activity during enzyme
assay. We could not detect any oxidation of 14,15-EET.
Microsomes of CYP4F8 oxidized 8,9-EET and 11,12-EET to their
corresponding 18-hydroxy metabolites (Fig. 7). This was confirmed
by a characteristic signal at m/z 277 (Aꢀ – 58; loss of CH3–CH2–

T. Nilsson et al. / Archives of Biochemistry and Biophysics 494 (2010) 64–71
69
100
100
80
60
40
20
0
A
B
TIC
m/z 179
80
60
40
20
0
6
9
12
15
5
7
9
11
13
Retention Time (min)
215
317.3
215
317.3
-CO
-H O
100
80
60
40
20
0
C
D
2
2
100
-CO
-H O
2
2
OH
277
+H
OH
277
179
O
167
-H
80
60
-
OOC
+H
-
167
OOC
167.2
O
155
179.2
40
20
0
273.3
277.3
277.3
273.4
255.3
215.3
155.2
167.2
255.3
299.3
299.3
215.3
123.2
120
135.3
140
240
280
320
160
200
100
180
220
m/z
260
300
340
m/z
Fig. 7. RP-HPLC analysis of metabolites formed by recombinant CYP4F8. (A) Partial total ion current chromatogram (TIC) showing the elution of a metabolite formed by
microsomes of CYP4F8 and 8,9-EET. The MS/MS spectrum is shown in C. (B) Partial ion chromatogram (m/z 179) showing the elution of a metabolite formed by microsomes of
CYP4F8 and 11,12-EET. The MS/MS spectrum is shown in (D). (C and D) The MS/MS spectra (m/z 335 ? full scan) of the main products formed from 8,9-EET and 11,12-EET
showed a characteristic ion at m/z 277 (335 ꢀ 58), and were identified as the
x
3 hydroxy metabolites (cf. Ref. [38]).
Oxidation of HEETs
somes was too low for quantification by the absorbance at
450 nm after Na2S2O4 reduction and treatment with CO.
Microsomes of CYP4F8 oxidized 10,11R,12R-HEET to the 18-hy-
droxy metabolite (Fig. 8A). The MS/MS spectrum (m/z 351 ? full
scan) showed a characteristic signal at m/z 293 (loss of O@CH–CH2–
CH3), m/z 231 (293-44-18, loss of CO2+ H2O), and m/z 213
(231 ꢀ 18) (Fig. 8B). Recombinant CYP4F8 did not oxidize 8(R/
S),11R,12R-HEET, 11S,14S,15S-HEET, 11R,14S,15S-HEET, or 13S,14S,-
15S-HEET significantly (data not shown).
Oxidation by recombinant CYP4F22
20:4nꢀ6
Microsomes of CYP4F22 were incubated with 200
l
M 20:4nꢀ6.
LC–MS/MS analysis (m/z 319 ? full scan) demonstrated the forma-
tion of two products, 16-hydroxyeicosa-5(Z),8(Z),11(Z),14(Z)-tetra-
enoic acid (16-HETE) (Fig. 10, top) and 18-hydroxyeicosa-
5(Z),8(Z),11(Z),14(Z)-tetraenoic acid (18-HETE) (Fig. 10, bottom).
Addition of cytochrome b5 did not increase the amounts of prod-
ucts formed (data not shown). The MS/MS spectrum of 18-HETE
Expression of recombinant CYP4F22 in yeast
The sequence of the amplified CYP4F22 open reading frame was
in accordance with the previously published sequence (GenBank
Accession No. NM_173483), except in position 582, where A was
was as reported [5], with a characteristic signal at m/z 261 (Aꢀ
–
exchanged for
G
(a reported synonymous SNP; GenBank,
58; loss of CH3–CH2–CHO). 16-HETE had a characteristic signal at
m/z 233 (Aꢀ – 86; loss of CH3–(CH2)3–CHO). Control experiments
with microsomes of yeast transformed with plasmid vector with-
out the CYP4F22 sequence, and microsomes of untransformed
yeast revealed the formation of rac 16-HETE (insert top Fig. 10),
which was therefore not associated with expression of CYP4F22.
rs11666601). The CYP4F22 sequence was ligated with the vector
pYeDP60 and used to transform the yeast strain W(R). Transforma-
tion was confirmed by CYP4F22 specific PCR analysis of genomic
DNA, which was isolated from transformed W(R) cells (Fig. 9).
The expression of recombinant CYP4F22 (and CYP4F8) in micro-

70
T. Nilsson et al. / Archives of Biochemistry and Biophysics 494 (2010) 64–71
333.3
100
100
80
A
B
231
I
m/z 195
-CO
-H2O
2
OH
80
O
293
-
OOC
60
40
60
OH
40
20
0
307.4
271.4
293.4
195.3
20
181.3
231.3
213.4
257.3
153.3
0
22
20
Retention Time (min)
24
18
120
180
240
m/z
300
360
Fig. 8. LC–MS/MS analysis of products formed by recombinant CYP4F8 and 10,11,12-HEET. (A) NP-HPLC analysis shows separation of products (hexane/isopropanol/acetic
acid, 95/5/0.01) with MS/MS analysis (m/z 351 ? full scan). (B) The MS/MS spectrum (m/z 351 ? full scan) of the material eluting at 21–22 min showed characteristic ions at
m/z 231, and m/z 213, consistent with 18-hydroxy-10,11,12-HEET.
100
R
m/z 233
S
80
60
40
20
16-HETE
Fig. 9. PCR analysis of transformation of yeast W(R) with CYP4F22 analyzed by
agarose gel electrophoresis. (A) PCR analysis of genomic DNA from W(R)
transformed with vector with the CYP4F22 open reading frame. (B) PCR analysis
of genomic DNA from W(R) transformed with vector without insert. (C) PCR
analysis of genomic DNA of untransformed W(R). (D) PCR analysis without template
DNA. L, low MW ladder (New England Biolabs).
10
14
min
100
80
60
40
20
0
Furthermore, the formation was NADPH dependent. We conclude
that 16-HETE was likely formed by endogenous CYP of yeast as a
minor product. Importantly, 18-HETE was not detected in any of
these control experiments, and 18-HETE can not be formed non-
enzymatically. We conclude that this metabolite was formed by
CYP4F22. In comparison with recombinant CYP4F8, the catalytic
activity of recombinant CYP4F22 was 1–2 orders of magnitude
lower.
m/z 261
18-HETE
HEETs
We could not detect any metabolites formed by recombinant
CYP4F22 in two experiments with 8,11R,12R-HEET or with
10,11R,12R-HEET. Under similar conditions, 10,11R,12R-HEET was
oxidized by CYP4F8, as described above.
6
7
8
Retention Time (min)
Discussion
Fig. 10. LC–MS/MS analysis of products formed from recombinant CYP4F22 and
20:4nꢀ6. Top, Selective ion monitoring of m/z 233 (319 ꢀ 86), a characteristic ion of
16-HETE. Bottom, selective ion monitoring of m/z 261 (319 ꢀ 58), a characteristic
ion of 18-HETE. The insert in the top chromatogram shows CP-HPLC analysis of 16-
HETE (16R-HETE above and the racemic biological sample below).
We report expression of recombinant CYP4F22 with arachido-
nate
x3 hydroxylase activity. This confirms that the open reading
frame of CYP4F22 codes for a functional enzyme, but unfortunately
the enzyme appeared to be poorly expressed. As we expected that
CYP4F22 also could metabolize HEETs, we performed a systematic
study of the MS/MS spectra of HEETs in order to determine metab-
olites formed by CYP4F22 and by CYP4F8.
CYP4F8 may thus be functionally linked to COX-2 of seminal vesi-
cles rather than to phospholipases and release of 20:4nꢀ6. It is
conceivable that CYP4F22 could be linked to the 12R-LOX and
eLOX3 pathway in a similar way.
We report a systematic study of MS/MS spectra of HEETs de-
rived from 12- and 15-HPETE, the two main LOX products in skin.
The MS/MS spectra of the HEETs followed the fragmentation pat-
tern described for the corresponding epoxyalcohols of linoleic acid
[34]. The proposed fragmentation was supported by MS/MS spec-
tra of [2H8]labeled isotopomers. We found that NP-HPLC–MS/MS
18-HETE cannot be formed non-enzymatically. Biosynthesis of
this metabolite by recombinant CYP4F22 is the first reported cata-
lytic activity associated with this enzyme. 18-HETE is also formed
by CYP4F8 [5], but the biological significance of this product is un-
known. Seminal fluid contains ꢂ0.1 mM concentration of 19-hy-
droxy-PGs, which are formed by
x2 hydroxylation of PGH by
CYP4F8 [5], but 18-HETE has not been detected in this fluid.

T. Nilsson et al. / Archives of Biochemistry and Biophysics 494 (2010) 64–71
71
[5] J. Bylund, M. Hidestrand, M. Ingelman-Sundberg, E.H. Oliw, J. Biol. Chem. 275
(2000) 21844–21849.
[6] F.P. Guengerich, Z.L. Wu, C.J. Bartleson, Biochem. Biophys. Res. Commun. 338
(2005) 465–469.
[7] M.J. Holtzman, J. Turk, A. Pentland, J. Clin. Invest. 84 (1989) 1446–1453.
[8] K. Stark, H. Törmä, M. Cristea, E.H. Oliw, Arch. Biochem. Biophys. 409 (2003)
188–196.
[9] C. Lefèvre, B. Bouadjar, V. Ferrand, G. Tadini, A. Megarbane, M. Lathrop, J.F.
Prud’homme, J. Fischer, Hum. Mol. Genet. 15 (2006) 767–776.
[10] W.E. Boeglin, R.B. Kim, A.R. Brash, Proc. Natl. Acad. Sci. USA 95 (1998) 6744–
6749.
[11] P.M. Woollard, Biochem. Biophys. Res. Commun. 136 (1986) 169–176.
[12] N. Epp, G. Fürstenberger, K. Müller, S. de Juanes, M. Leitges, I. Hausser,
F. Thieme, G. Liebisch, G. Schmitz, P. Krieg, J. Cell Biol. 177 (2007)
173–182.
[13] G. Fürstenberger, N. Epp, K.M. Eckl, H.C. Hennies, C. Jorgensen, P. Hallenborg, K.
Kristiansen, P. Krieg, Prostaglandins Other Lipid Mediat. 82 (2007) 128–134.
[14] J.L. Moran, H. Qiu, A. Turbe-Doan, Y. Yun, W.E. Boeglin, A.R. Brash, D.R. Beier,
J. Invest. Dermatol. 127 (2007) 1893–1897.
[15] F. Jobard, C. Lefèvre, A. Karaduman, C. Blanchet-Bardon, S. Emre,
J. Weissenbach, M. Özgüc, M. Lathrop, J.F. Prud’homme, J. Fischer, Hum. Mol.
Genet. 11 (2002) 107–113.
[16] K.M. Eckl, S. de Juanes, J. Kurtenbach, M. Natebus, J. Lugassy, V. Oji, H. Traupe,
M.L. Preil, F. Martinez, J. Smolle, A. Harel, P. Krieg, E. Sprecher, H.C. Hennies,
J. Invest. Dermatol. 129 (2009) 1421–1428.
analysis was useful for analysis of HEETs in biological samples, as
illustrated in human corneal tissue.
8R,11R,12R-HEET is produced in human epidermis by 12R-LOX
and eLOX3 [19]. A linkage between mutations in the gene of
CYP4F22 and lamellar ichtyosis was reported by Lefèvre et al.,
and the authors proposed that CYP4F22 could be involved in oxida-
tion of 8R,11R,12R-HEET [9]. However, neither CYP4F22 nor
CYP4F8 appeared to oxidize 8,11R,12R-HEET. Recombinant CYP4F8
oxidized 10,11R,12R-HEET to its 18-hydroxy metabolite (Fig. 8A),
but 11S,14S,15S-HEET, 11R,14S,15S-HEET and 13S,14S,15S-HEET
were not substrates.
The selective oxygenation of 10,11R,12R-HEET is interesting, as
both 12R-LOX and CYP4F8 are up regulated in psoriatic lesions
[8,11]. To investigate whether the position of the epoxide group
at C11/C12 contributed to this selectivity, we tested the oxygena-
tion of EETs. We found that 8,9-EET and 11,12-EET were oxidized
at the
x3 position, but 14,15-EET was not a substrate. The position
of the epoxide group was apparently important. No significant
non-enzymatic hydrolysis of the HEETs or EETs was observed in
our enzyme assays.
HEETs can be hydrolyzed to triols by epoxide hydrolases in tis-
sues, and these products might be substrates of CY4F members. It
is therefore possible that CYP4F22 might be involved in a subse-
quent step in the 12R-LOX/eLOX3 pathway. However, CYP4F22 is
also expressed in testes, where 12R-LOX has not been described.
It is conceivable that CYP4F22 could be involved in oxygenation
[17] Z. Yu, C. Schneider, W.E. Boeglin, L.J. Marnett, A.R. Brash, Proc. Natl. Acad. Sci.
USA 100 (2003) 9162–9167.
[18] F. Lesueur, B. Bouadjar, C. Lefèvre, F. Jobard, S. Audebert, H. Lakhdar, L. Martin,
G. Tadini, A. Karaduman, S. Emre, S. Saker, M. Lathrop, J. Fischer, J. Invest.
Dermatol. 127 (2007) 829–834.
[19] Z. Yu, C. Schneider, W.E. Boeglin, A.R. Brash, Biochim. Biophys. Acta 1686 (3)
(2005) 238–247.
[20] K.M. Eckl, P. Krieg, W. Küster, H. Traupe, F. André, N. Wittstruck,
G. Fürstenberger, H.C. Hennies, Hum. Mutat. 26 (2005) 351–361.
[21] C. Lefèvre, B. Bouadjar, A. Karaduman, F. Jobard, S. Saker, M. Ozguc, M. Lathrop,
J.F. Prud’homme, J. Fischer, Hum. Mol. Genet. 13 (2004) 2473–2482.
[22] C.R. Pace-Asciak, Biochim. Biophys. Acta 1215 (1994) 1–8.
[23] S. Nigam, M.P. Zafiriou, R. Deva, R. Ciccoli, R. Roux-Van der Merwe, FEBS J. 274
(2007) 3503–3512.
[24] A.A. Spector, J. Lipid Res. 50 (Suppl.) (2009) S52–S56.
[25] P.A. Ladd, L. Du, J.H. Capdevila, R. Mernaugh, D.S. Keeney, J. Biol. Chem. 278
(2003) 35184–35192.
[26] Z. Yu, C. Schneider, W.E. Boeglin, A.R. Brash, Lipids 42 (2007) 491–497.
[27] L.A. Cowart, S. Wei, M.-H. Hsu, E.F. Johnson, M.U. Krishna, J.R. Falck, J.H.
Capdevila, J. Biol. Chem. 277 (2002) 35105–35112.
[28] J. Bylund, N. Finnstrom, E.H. Oliw, Biochem. Biophys. Res. Commun. 261 (1999)
169–174.
of other lipids in analogy with CYP4F8 and
x2 hydroxylation of
PGH in seminal vesicles [5]. Further studies of the function of
CYP4F22 in reproductive organs seem merited.
In summary, we report that recombinant CYP4F22 catalyzed
x3
hydroxylation of 20:4nꢀ6, whereas oxygenation of 8,11R,12R-
HEET was not detected. CYP4F8 oxidized 8,9-EET, 11,12-EET, and
10,11R,12R-HEET. The latter is formed from 12R-HPETE, which
suggests a functional link between CYP4F8 and 12R-LOX in
epidermis.
[29] K. Stark, B. Wongsud, R. Burman, E.H. Oliw, Arch. Biochem. Biophys. 441 (2005)
174–181.
Acknowledgments
[30] H. Yin, N.A. Porter, Antioxid. Redox Signal. 7 (2005) 170–184.
[31] I.V. Ivanov, S.G. Romanov, N.V. Groza, S. Nigam, H. Kuhn, G.I. Myagkova, Bioorg.
Med. Chem. 10 (2002) 2335–2343.
[32] S. Harju, H. Fedosyuk, K.R. Peterson, BMC Biotechnol. 4 (2004) 8.
[33] T.A. Dix, L.J. Marnett, J. Biol. Chem. 260 (1985) 5351–5357.
[34] E.H. Oliw, U. Garscha, T. Nilsson, M. Cristea, Anal. Biochem. 354 (2006) 111–
126.
[35] W.C. Chang, J. Nakao, H. Orimo, S. Murota, Biochem. J. 202 (1982) 771–776.
[36] E.H. Oliw, J. Chromatogr. 275 (1983) 245–259.
[37] U. Garscha, T. Nilsson, E.H. Oliw, J. Chromatogr, B. Analyt, Technol. Biomed. Life
Sci. 872 (2008) 90–98.
[38] T. Nakamura, D.L. Bratton, R.C. Murphy, J. Mass Spectrom. 32 (1997)
888–896.
[39] J. Bylund, J. Ericsson, E.H. Oliw, Anal. Biochem. 265 (1998) 55–68.
[40] R.M. Hanson, in: L.E. Overman (Ed.), Organic Reactions, John Wiley & Sons Inc.,
New York, 2002, pp. 1–157.
This work was supported by Vetenskapsrådet Medicin (3X-
06523), Stiftelsen Lars Hiertas Minne (KDB269/08), and The Knut
and Alice Wallenberg Foundation (2004.0123).
References
[1] D.R. Nelson, D.C. Zeldin, S.M. Hoffman, L.J. Maltais, H.M. Wain, D.W. Nebert,
Pharmacogenetics 14 (2004) 1–18.
[2] M. Fer, L. Corcos, Y. Dreano, E. Plee-Gautier, J.-P. Salaun, F. Berthou, Y. Amet, J.
Lipid Res. 49 (2008) 2379–2389.
[3] Y. Kikuta, M. Kato, Y. Yamashita, Y. Miyauchi, K. Tanaka, N. Kamada, M.
Kusunose, DNA Cell Biol. 17 (1998) 221–230.
[4] Y. Kikuta, Y. Miyauchi, E. Kusunose, M. Kusunose, DNA Cell Biol. 18 (1999)
723–730.