Novel insights into cyclooxygenases, linoleate diol synthases, and lipoxygenases from deuterium kinetic isotope effects and oxidation of substrate analogs

Article abstract of DOI:10.1016/j.bbalip.2012.09.001

Cyclooxygenases (COX) and 8R-dioxygenase (8R-DOX) activities of linoleate diol synthases (LDS) are homologous heme-dependent enzymes that oxygenate fatty acids by a tyrosyl radical-mediated hydrogen abstraction and antarafacial insertion of O₂. Soybean lipoxygenase-1 (sLOX-1) contains non-heme iron and oxidizes 18:2n - 6 with a large deuterium kinetic isotope effect (D-KIE). The aim of the present work was to obtain further mechanistic insight into the action of these enzymes by using a series of n - 6 and n - 9 fatty acids and by analysis of D-KIE. COX-1 oxidized C₂₀ and C₁₈ fatty acids in the following order of rates: 20:2n - 6 > 20:1n - 6 > 20:3n - 9 > 20:1n - 9 and 18:3n - 3 ≥ 18:2n - 6 > 18:1n - 6. 18:2n - 6 and its geometrical isomer (9E,12Z)18:2 were both mainly oxygenated at C-9 by COX-1, but the 9Z,12E isomer was mostly oxygenated at C-13. A cis-configured double bond in the n - 6 position therefore seems important for substrate positioning. 8R-DOX oxidized (9Z,12E)18:2 at C-8 in analogy with 18:2n - 6, but the 9E,12Z isomer was mainly subject to hydrogen abstraction at C-11 and oxygen insertion at C-9 by 8R-DOX of 5,8-LDS. sLOX-1 and 13R-MnLOX oxidized [11S- ²H]18:2n - 6 with similar D-KIE (～ 53), which implies that the catalytic metals did not alter the D-KIE. Oxygenation of 18:2n - 6 by COX-1 and COX-2 took place with a D-KIE of 3-5 as probed by incubations of [11,11- ²H₂]- and [11S-²H]18:2n - 6. In contrast, the more energetically demanding hydrogen abstractions of the allylic carbons of 20:1n - 6 by COX-1 and 18:1n - 9 by 8R-DOX were both accompanied by large D-KIE (> 20).

Full text of DOI:10.1016/j.bbalip.2012.09.001

Biochimica et Biophysica Acta 1821 (2012) 1508–1517
Contents lists available at SciVerse ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbalip
Novel insights into cyclooxygenases, linoleate diol synthases, and lipoxygenases from
deuterium kinetic isotope effects and oxidation of substrate analogs
Inga Hoffmann a, Mats Hamberg b, Roland Lindh c, Ernst H. Oliw a,
⁎
a
Division of Biochemical Pharmacology, Department of Pharmaceutical Biosciences, Uppsala University, Biomedical Center, SE-751 24 Uppsala, Sweden
Division of Physiological Chemistry II, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, SE-171 77 Solna, Sweden
Department of Chemistry - Ångström, The Theoretical Chemistry Programme, Uppsala University, SE-751 20 Uppsala, Sweden
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 25 April 2012
Received in revised form 20 August 2012
Accepted 4 September 2012
Available online 12 September 2012
Cyclooxygenases (COX) and 8R-dioxygenase (8R-DOX) activities of linoleate diol synthases (LDS) are homologous
heme-dependent enzymes that oxygenate fatty acids by a tyrosyl radical-mediated hydrogen abstraction and
antarafacial insertion of O2. Soybean lipoxygenase-1 (sLOX-1) contains non-heme iron and oxidizes 18:2n−6
with a large deuterium kinetic isotope effect (D-KIE). The aim of the present work was to obtain further mechanis-
tic insight into the action of these enzymes by using a series of n−6 and n−9 fatty acids and by analysis of D-KIE.
COX-1 oxidized C20 and C18 fatty acids in the following order of rates: 20:2n−6>20:1n−6>20:3n−9>20:1n−9
and 18:3n−3≥18:2n−6>18:1n−6. 18:2n−6 and its geometrical isomer (9E,12Z)18:2 were both mainly oxy-
genated at C-9 by COX-1, but the 9Z,12E isomer was mostly oxygenated at C-13. A cis-configured double bond in
the n−6 position therefore seems important for substrate positioning. 8R-DOX oxidized (9Z,12E)18:2 at C-8 in anal-
ogy with 18:2n−6, but the 9E,12Z isomer was mainly subject to hydrogen abstraction at C-11 and oxygen insertion
at C-9 by 8R-DOX of 5,8-LDS. sLOX-1 and 13R-MnLOX oxidized [11S-2H]18:2n−6 with similar D-KIE (~53), which
implies that the catalytic metals did not alter the D-KIE. Oxygenation of 18:2n−6 by COX-1 and COX-2 took place
with a D-KIE of 3–5 as probed by incubations of [11,11-2H2]- and [11S-2H]18:2n−6. In contrast, the more energet-
ically demanding hydrogen abstractions of the allylic carbons of 20:1n−6 by COX-1 and 18:1n−9 by 8R-DOX were
both accompanied by large D-KIE (>20).
Keywords:
Animal heme peroxidase
Chiral phase HPLC
Fatty acid oxygenation
Kinetic isotope effect
Mass spectrometry
Oxygenation mechanism
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
function of eicosanoids has inspired the development of enzyme inhibi-
tors and receptor agonists and antagonists [3,4]. In plants and fungi, the
Polyunsaturated fatty acids can be oxygenated to important biological
mediators. The first steps are usually catalyzed by fatty acid dioxygenases
(DOX), which include heme-dependent DOX and lipoxygenases (LOX), or
by monooxygenases (cytochromes P450) [1–4]. DOX contain heme and
LOX non-heme iron (or rarely manganese). The DOX prototype is prosta-
glandin H (PGH) synthase-1, often designated cyclooxygenase (COX-1),
and the most thoroughly studied LOX are soybean LOX-1 (sLOX-1), and
human 5-LOX. In humans, COX and 5-LOX oxidize 20:4n−6 to
prostaglandins and leukotrienes, respectively, which can activate
G-protein coupled receptors in control of physiological functions,
fever, inflammation, and development of cancer [3,4]. The versatile
products of fatty acid DOX pathways, often referred to as oxylipins, partic-
ipate in plant defence and fungal development and pathogenicity [2,5–7].
These oxylipins are formed by LOX or by heme-dependent DOX, viz.,
α-DOX of plants, and linoleate diol synthases (LDS) and related enzymes
of fungi [2,6,8].
The LOX family constitutes homologous enzymes with a single
peptide chain, folded into a small N-terminal β-barrel domain and a
C-terminal domain with the catalytic metal [1,9]. All LOX contain cata-
lytic iron except two enzymes with manganese [10,11]. The metal
center redox cycles in a process of proton-coupled electron transfer
from the C\H bond of a bis-allylic carbon to the catalytic base of the
metal center (Fe3+OH−), followed by antarafacial insertion of molecu-
lar oxygen [1]. The hydrogen abstraction is characterized by a large deu-
terium kinetic isotope effect (D-KIE) of ~80 [12]. Whether the D-KIE
differs between Fe- and MnLOX is unknown. Differences in zero point
energies and tunneling contribute to the D-KIE [13]. The zero point en-
ergy accounts for the 5- to 10-fold reduction of the reaction rate, which
is due to the difference of activation energy needed for breaking C\H
and C\D bonds (~1 kcal/mol). Larger D-KIE (>20) can be explained
by hydrogen tunneling through the energy barrier, and this has been
demonstrated for sLOX-1 by Klinman and co-workers [12,14]. The
Abbreviations: CP, chiral phase; COX, cyclooxygenase; DOX, dioxygenase;
HODE, hydroxyoctadecadienoic acid; HETE, hydroxyeicosatetraenoic acid;
HPEME, hydroperoxyeicosenoic acid; HPOME, hydroperoxyoctadecenoic acid;
HEME, hydroxyeicosenoic acid; HOME, hydroxyoctadecenoic acid; D-KIE, deu-
terium kinetic isotope effect; LC, liquid chromatography; LDS, linoleate diol
synthase; LOX, lipoxygenase; MS, mass spectrometry; NP, normal phase;
[11R-2H] (9Z,12E)18:2, [11R-2H]-9Z,12E-octadecadienoic acid; PGH, prostaglan-
din H; RP, reversed phase; RSV, ram seminal vesicles; sLOX, soybean
lipoxygenase; TIC, total ion current
⁎
Corresponding author. Tel.: +46 18 471 44 55; fax: +46 184714847.
E-mail address: Ernst.Oliw@farmbio.uu.se (E.H. Oliw).
1388-1981/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.bbalip.2012.09.001

I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
1509
extent of tunneling during sLOX-1 catalysis is influenced by structural
2.2. Synthesis of mono- and dideuterated fatty acids
fluctuations and dynamic effects, as shown by these authors [14].
The heme-dependent DOX are homologous to animal heme peroxi-
dases, and share fundamental catalytic properties [5,8,15]. Oxidation of
the heme group of COX, α-DOX, and LDS generates in a peroxidase
cycle a tyrosyl radical (Tyr•), which catalyzes hydrogen removal. Tyr• of
COX-1 abstracts the proS hydrogen at C-13 of 20:4n−6, and this occurs
with a D-KIE of ~2–3 at 20–30 °C [1,4,16], whereas Tyr• of α-DOX
abstracts hydrogen of palmitate with a D-KIE of ~54 [17]. A large D-KIE
(>20) was also reported for COX-2 with perdeuterated 18:2n−6 as a
substrate [18]. 7,8- And 5,8-LDS are fusion proteins of N-terminal
domains with 8R-DOX activities and C-terminal P450 domains with diol
synthase activities [19,20]. The Tyr• of 8R-DOX of 7,8- and 5,8-LDS
abstracts the proS hydrogen from the allylic position at C-8 of 18:2n−6
and 18:1n−9 [5,15,21], whereas the heme-thiolate abstracts the proS
hydrogens at C-7 and C-5, respectively [5,21]. Recently, 5,8-LDS of
Aspergillus nidulans1 was found to catalyze these oxidations at C-8 and
at C-5 of 18:1n−9 with D-KIE ~33 and ~1.1, respectively [22].
sLOX-1 and COX-1 oxidize 20:4n−6 with more than a 10-fold differ-
ence in D-KIE. This large difference is enigmatic, but could be due to kinet-
ic factors, variable transitional complexes, and hydrogen donor–acceptor
distances [16]. The C\H bond dissociation enthalpies of bis-allylic and
allylic carbons are ~73 and ~87 kcal/mol, respectively [23], and this differ-
ence may influence the magnitude of the D-KIE. Suboptimal substrates
may have higher D-KIE than native substrates due to steric factors.
The catalytic base of sLOX-1, Fe3+OH−, is restricted to hydrogen
abstraction from bis-allylic carbons [24], whereas Tyr• of both COX-1
and LDS can oxidize allylic carbons [25]. Based on the large D-KIE of
LOX, we hypothesized that a larger D-KIE of COX-1 and LDS might be
observed during activation of C\H at singly allylic than bis-allylic posi-
tions since the contribution of hydrogen tunneling to the reaction rate is
likely larger in reactions, which require more energy.
2.2.1. [11,11-2H2]18:2n−6
2-Octynoic acid (9.2 g; Sigma-Aldrich) was refluxed for 1 h with
200 ml of methanol containing 2 ml conc. hydrochloric acid. The
resulting methyl ester (8.9 g) was dissolved in 30 ml of diethyl ether
and slowly added at 0 °C under magnetic stirring to a suspension of
1.8 g LiAlD4 in 50 ml of diethyl ether. After stirring at 0 °C for 1 h, the re-
action was quenched by slow addition of tetrahydrofuran/water and the
alcohol (7.8 g) was isolated by extraction with diethyl ether. The material
was dissolved in 50 ml of diethyl ether containing 1 ml of dry pyridine.
PBr3 (8.1 g) was slowly added at 0 °C under magnetic stirring, and the so-
lution was subsequently refluxed for 1.5 h. [1,1-2H2]1-bromo-2-octyne
(7.9 g; yield, 63%) was obtained following purification on a silica gel col-
umn (elution with hexane). [1,1-2H2]1-bromo-2-octyne (1.91 g) and
methyl 9-decynoate (1.82 g) were stirred at 23 °C for 3 h with 1 eq. of
CuI, 2 eq. of NaI, and 2 eq. of Cs2CO3, suspended in 30 ml of dry N,
N-dimethylformamide (cf. [28]). After quenching with ammonium
chloride, extraction with diethyl ether, drying over MgSO4 and chroma-
tography on silica gel, the methyl ester of the title compound was
obtained in >90% yield. Part of this material (584 mg) was subjected to
semihydrogenation using P-2 nickel [29] and purified by RP-HPLC. An
aliquot of the deuterated methyl 9Z,12Z-octadecadienoate (184 mg)
was saponified, and the free acid purified by RP-HPLC. This afforded the
title compound (89 mg) as a colorless oil (purity >98%). The isotope
composition as determined by GC–MS was 94.3% dideuterated, 5.1%
monodeuterated, and 0.6% undeuterated molecules.
2.2.2. [8,8-2H2]18:1n−9
8-Hydroxyoctanoic acid, prepared by alkali treatment of 8-
bromooctanoic acid (Sigma-Aldrich), was treated with dihydro
pyran/p-toluenesulfonic acid to provide the tetrahydropyranyl ester/
ether derivative. This material (5.99 g) in 100 ml of THF was
refluxed with LiAlD4 (3 g) for 2 h, and the deuterated 8-
tetrahydropyranyloxy-1-octanol (4.47 g) was sequentially treat-
ed with methanesulfonyl chloride/triethylamine and sodium
iodide to afford 8-tetrahydropyranyloxy-1-iodooctane (3.63 g;
yield, 58%). An aliquot of this material (1 g; 2.9 mmol) in 5 ml
of THF was added to the lithio derivative of 1-decyne (9 mmol)
in 15 ml of THF and 5 ml of DMPU at 0 °C. The mixture
was stirred at 0 °C for 2.5 h, quenched with ammonium chloride
and extracted with diethyl ether. The resulting deuterated 1-
tetrahydropyranyloxy-9-octadecyne was deprotected by treatment with
p-toluenesulfonic acid, and the deuterated acetylenic alcohol (0.65 g;
2.43 mmol) was oxidized with pyridinium dichromate in N,N-
dimethylformamide [30]. Semihydrogenation of the methyl ester using
P-2 nickel [29], saponification, and purification by RP-HPLC afforded the
title compound as a colorless oil (0.26 g; purity, >98%; >96% deuterated
molecules; yield from the deuterated acetylenic alcohol, 38%).
The first goal was to investigate mono- and polyunsaturated fatty
acids and stereoisomers of 18:2n−6 as substrates of COX-1, 8R-DOX,
and sLOX-1. The second goal was to determine the D-KIE during hydro-
gen abstraction and oxidation of bis-allylic carbons by COX-1 and
COX-2, and allylic carbons by COX-1 and 8R-DOX. The third goal was to
compare the D-KIE of LOX with Fe3+OH− and Mn3+OH− as catalytic
bases.
2. Materials and methods
2.1. Materials
HPLC solvents (Lichrosolve) and routine chemicals were from Merck.
18:1n−9 (99%), 20:1n−6 (99%), 20:1n−9 (99%), 20:4n−6 (99%), and
13
[
C ]18:2n−6 were from Larodan. 18:1n−6 (99%), (9E,12Z)- and
18
(9Z,12E)-18:2, 9R,S-HODE(10E,12E) and 9S-HODE(10E,12E) were from
Lipidox. [11S-2H]18:2n−6 was prepared as described [5,26]. 20:3n−9
(99%) and purified ovine COX-2 (4.1 kU/mg) were from Cayman.
18:3n−6, 18:2n−6, (9E)-18:1, hematin, N-hydroxyphthalimide, ceric
ammonium nitrate, lipoxidase type 5 (sLOX-1), and α-tocopherol were
from Sigma-Aldrich. Fatty acids were dissolved in ethanol and stored in
2.2.3. [13,13-2H2]20:1n−6
12-Bromo-1-dodecanol (1 g; Sigma-Aldrich) was treated with
dihydropyran/p-toluenesulfonic acid to afford 1-bromo-12-tetra-
hydropyranyloxy-dodecane (1.3 g; yield, 99%). The Grignard derivative
of this material (3.7 mmol) in 5 ml of THF was stirred at 23 °C
for 5 min with 60 μmol of dilithium tetrachlorocuprate(II) and then
stirred for further 3 h following addition of 764 mg (4 mmol) of
[1,1-2H2]1-bromo-2-octyne. Extractive isolation with diethyl ether, re-
moval of the tetrahydropyranyl blocking group by treatment with p-
toluenesulfonic acid and chromatography on a silica gel column afforded
material (0.54 g) consisting of the desired deuterated 14-eicosyn-1-ol
admixed with a branched chain rearrangement product, ratio ~1:2. An
aliquot of the mixture was oxidized by treatment with pyridinium dichro-
mate in N,N-dimethylformamide [30], and pure deuterated 14-eicosynoic
acid (70 mg) was obtained following RP-HPLC. Semihydrogenation of the
methyl ester using P-2 nickel [29], saponification to the free acid, and
stock solutions (30–100 mM) at −20 °C. Photooxidation of [13C18
]
18:2n−6 was performed with methylene blue in methanol. Microsomes
of ram seminal vesicles (RSV) were prepared as described [27]. Recombi-
nant 8R-DOX domains (5,8-LDS, residues 1–674 of Aspergillus fumigatus;
7,8-LDS, residues 1–673 of Gaeumannomyces graminis) were expressed
in Escherichia coli [20], and referred to as 8R-DOX of 5,8- and 7,8-LDS.
Full-length 5,8- and 7,8-LDS were expressed in E. coli [20]. Chiralcel
OB-H (250×4.6 mm; Daicel) was purchased locally (Dalco Chromtech).
Recombinant 13R-MnLOX was prepared as described [24].
1
5,8-LDS is also designated psi producing oxygenase A (ppoA) [19].

1510
I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
purification by RP-HPLC afforded the title compound (45 mg; purity
2.6. Assays of non-competitive D-KIE
>98%; >96% deuterated molecules).
2.6.1. COX-1
2.2.4. [11R-2H] (9Z,12E)-18:2
We first compared the oxygen consumptions of COX-1 with
18:2n−6, [11S-2H]18:2, [11,11-2H2]18:2n−6, 20:1n−6 and
[13,13-2H2]20:1n−6 as substrates. As the D-KIE of the latter was
too large for quantification, we assayed the non-competitive D-KIE
with LC–MS/MS. Microsomes of RSV (COX-1) were incubated in the
same way as for analysis of oxygen consumption above (triplicate
at 30 °C for 150 s) with 100 μM 20:1n−6 and with 100 μM
[13,13-2H2]20:1n−6. The reactions were ended by addition of meth-
anol with the internal standard (0.3 μg 13-HODE), hydroperoxides
were reduced by NaBH4, and the products were extracted as above
and analyzed by LC–MS/MS (13-HODE, m/z 195; 13- and 15-HEME,
m/z 225 and 227; [2H]13- and [2H]15-HEME, m/z 226 and 228).
A recently described method was used [26]. Briefly, the epoxy alcohol
methyl 11R,12R-epoxy-13S-hydroxy-9Z-octadecenoate (300 mg) was
stirred with LiAlD4 in diethyl ether affording trideuterated 12S,13S-
dihydroxy-9Z-octadecenol. The threo-triol was sequentially converted
into the title compound following monoacetylation of the primary alcohol
function, conversion into the cyclic thionocarbonate derivative, stereo-
specific deoxygenation of the thionocarbonate, oxidation with pyridinium
dichromate in N,N-dimethylformamide, and purification by RP-HPLC. The
material obtained (purity >98%) was free from the 9Z,12Z-, 9E,12Z- and
9E,12E-octadecadienoate geometrical isomers as checked by GLC analysis
of the methyl ester.
2.3. Autoxidation of 20:1n−6 and [13,13-2H2]20:1n−6
2.6.2. COX-2
We compared the rate of oxidation of 100 μM 18:2n−6, 100 μM
[11S-2H]18:2, and 100 μM [11,11-2H2]18:2n−6 by COX-2 (12 μg) in
triplicates in 0.5–1 ml 0.05 M Tris–HCl (pH 8.0)/5 mM EDTA/1 μM hema-
tin (30 °C; 3 min). The reactions were initiated by the addition of COX-2
and terminated by methanol. 13-HOTrE (125 ng) was added as an inter-
nal standard and hydroperoxides were reduced to alcohols by NaBH4 in
methanol (+4 °C). The main products, 9-HODE and [11R-2H]9-HODE
(cf. [31]), respectively, were analyzed after extractive isolation as above
by LC–MS/MS with monitoring of the internal standard (m/z 195) and
m/z 171 from MS/MS analysis of m/z 295 and 296, respectively.
5 mg 20:1n−6 and 0.2 mg α-tocopherol were treated with N-
hydroxyphthalimide (0.8 eq.) and ceric ammonium nitrate (0.2 eq.) in
acetonitrile (0.2 ml) under oxygen at 37 °C. [13,13-2H2]20:1n−6
(5 mg) was oxidized in the same way. The reaction was followed by
TLC and terminated by extractive isolation after 3–4 days, when >50%
of 20:1n−6 was consumed. The products were separated by RP-HPLC
(methanol/acetonitrile/water/acetic acid, 450/310/240/0.2) in three
groups (13-HPEME, 16-HPEME, and mixture of 14- and 15-HPEME).
2.4. COX-1 activity assay by oxygen consumption
A YSI model 5300 biological oxygen monitor was used with a jacketed
cell (1.5 ml; OX-15253, Gilson), an YSI 5331 oxygen probe, standard
sensitivity membranes, and a magnetic stirrer. The cell was thermostated
at 30 °C. Pre-warmed buffer (0.05 M Tris–HCl (pH 8.0 at 25 °C)/1 μM he-
matin) at 30 °C and suspended microsomes of RSV on ice were added to
the cell, and allowed to equilibrate for at least 2 min. The reaction was ini-
tiated by addition of fatty acids in ethanol through the stopper to a final
concentration of 100 μM and oxygen consumption was followed for a
few minutes. The oxygen monitor was connected to an integrator
(Hitachi-Merck D-2500). Km of COX-1 and 20:1n−6 was estimated in
triplicates by non-linear regression to the Michaelis–Menten equation
with 10–300 μM 20:1n−6 (GraFit software).
2.6.3. 8R-DOX
Recombinant 8R-DOX of 5,8- and 7,8-LDS were incubated with
100 μM 18:1n−9 and [8,8-2H2]18:1n−9 in triplicate on ice and the reac-
tions were ended (after 1–3–10 min and 3–10–30 min, respectively) by
addition of methanol with the internal standard (0.3 μg 13-HODE).
Hydroperoxides were reduced with NaBH4 on ice, and the products
were isolated on a cartridge of C18 silica (SepPak/C18). For LC–MS/MS,
we monitored the internal standard at m/z 195, 8-HODE at m/z 157, and
[2H]8-HODE at m/z 158.
2.6.4. sLOX-1 and 13R-MnLOX
The sLOX-1 and 13R-MnLOX activities were assayed in 0.1 M NaBO3
buffer (pH 9.0; 23 °C) with UV analysis as described [24]. Dilutions of
sLOX-1 and 13R-MnLOX were first titrated to oxidize 18:2n−6 at virtual-
ly the same rate, and we then compared the rates of oxidation of 100 μM
[11S-2H]18:2n−6.
2.5. Enzyme assays by LC–MS and UV spectroscopy
2.5.1. COX-1
For structural analysis of products, microsomes of RSV were incubated
with 100 μM fatty acids in 0.1 M KHPO4 buffer (pH 7.4)/2 mM EDTA
for 30 min (37 °C). The reaction was stopped by addition of ethanol
(4 vols.) and proteins were precipitated by centrifugation. The superna-
tant was evaporated to dryness and the products were then extracted
(SepPak/C18). Fatty acid hydroperoxides were reduced in some experi-
ments to alcohols by treatment with triphenylphosphine or with NaBH4.
2.7. HPLC and MS analyses
RP-HPLC-MS/MS analysis was performed on an octadecasilyl silica
column (5 μm, 100 Å, 150×2 mm; Phenomenex), which was eluted at
0.3 ml/min with methanol/water/acetic acid, 80/20/0.01. CP-HPLC-MS/
MS analysis of hydroxyl fatty acids was performed on Chiralcel OB-H
(250×4.6 mm), eluted at 0.5 ml/min with hexane/isopropyl alcohol/
acetic acid, 95/5/0.01, or on Reprosil Chiral AM (250×2 mm), eluted at
0.2 ml min with hexane/ethanol/acetic acid, 95/5/0.01. The effluents
from the columns were combined with isopropyl alcohol/water (3/2)
from a second HPLC pump and then introduced by electrospray into a
linear ion trap mass spectrometer (LTQ, ThermoFisher). The transfer cap-
illary was heated to 315 °C, the ion isolation width was set at 1.5 for
anions of hydroxy fatty acids and 5 for anions of hydroperoxy fatty
acids and 1.5 at the final selection of MS3 analysis. The collision energy
was set at 1.7 V, and the ion tube lens at −110 to −130 V. We recorded
five microscans and used the Gaussian algorithm for peak smoothing
(Xcalibur software). Prostaglandin F1α was infused for tuning.
2.5.2. 8R-DOX
Recombinant 8R-DOX of 5,8- and 7,8-LDS were incubated in
0.05 M Tris–HCl (pH 7.6)/5 mM EDTA/10% glycerol with 100 μM
(9E,12Z)18:2, (9Z,12E)18:2, and (9E)18:1 for 30 min on ice, and the
products were extracted (SepPak/C18) for LC–MS analysis.
2.5.3. sLOX-1
sLOX-1 (4–7 μM) was incubated with 200 μM (9E,12Z)18:2 or
with 200 μM [11R-2H] (9Z,12E)18:2 in 0.1 M NaBO3 (pH 9.0; 23 °C),
and the reaction was followed by UV analysis (235 nm) as described
[24]. The products formed from 200 μM (9E,12Z)18:2 were analyzed
by chiral phase-HPLC (Reprosil Chiral AM) after addition of a small
amount of 9-and 13-[13C18]HODE.

I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
1511
2.8. Steric analysis after ozonolysis
indicating that the 13-peroxyl radical may not be an intermediate in
formation of 15-HPEME.
Steric analysis of HODE and HEME methyl esters was performed
by derivatization into (−)-menthoxycarbonyl derivatives followed
by GC–MS analysis of diastereomeric short chain fragments prepared
by oxidative ozonolysis and methylation [32].
3.2. Oxygenation of C18 and C20 fatty acids of the n−9 series by COX-1
18:1n−9 was not oxidized by COX-1 to a significant extent as judged
from oxygen consumption assay (Fig. 1A), but 20:1n−9 was oxidized,
although slowly compared to 20:1n−6 (Fig. 1B). The rate of oxidation
of 20:1n−9 increased gradually, and the rate of oxidation of Mead acid
(20:3n−9) increased in a similar way.
3. Results
3.1. Oxygenation of C18 and C20 fatty acids of the n−6 series by COX-1
From the initial rates of oxidation, the order of C20 fatty acids was
20:2n−6>20:1n−6>20:3n−9>20:1n−9, and the order of C18
fatty acids was 18:3n−3≥18:2n−6 >18:1n−6.
18:2n−6 binds to the active site of COX-1 in analogy with 20:4n−6
[33]. 18:2n−6 was rapidly oxidized by COX-1 [34], and so was 18:3n−6
(Fig. 1A). In analogy with 18:2n−6, 18:3n−6 was converted to 9-HOTrE
as the main metabolite and to small amounts of 13-HOTrE (LC–MS/
MS analysis; Fig. 2A). A minor polar metabolite yielded a MS/MS
spectrum, which was identical to the MS/MS spectrum of dinor-
PGE1 (Fig. 2B), reported by Murphy et al. [35]. 18:1n−6 was oxi-
dized at a slower rate than 18:3n−6 (Fig. 1A). Products formed
from 18:1n−6 were described previously (11-HOME(12Z) and
13-HOME(11E) in a ratio of ~4:1 [24]). Steric analysis by analysis of
chiral fragments demonstrated optically pure 11S-HOME, ~93%
13S-HOME and ~7% 13R-HOME.
20:1n−6 was oxidized to a mixture of 13-HEME(14Z) and 15-
HEME(13E) in a ratio of ~3:1 (Fig. 3A). Steric analysis showed that
13-HEME was formed with S configuration, and that 15-HEME consisted
of ~93% of the S stereoisomer. The Km was estimated to be 20 6 μM.
Racemic 13-HPEME was not transformed to 15-HPEME by COX-1,
3.3. Oxygenation of (9E,12Z)- and (9Z,12E)18:2 by COX-1
(9E,12Z)18:2 was oxidized by COX-1 to 9-HODE and (9E,11E)13-
HODE in a ratio of ~3:1. The absolute configuration was analyzed by
CP-HPLC (Fig. 4A) and by ozonolysis. The latter analysis showed the fol-
lowing composition of regio- and stereoisomers: 9R (8%), 9S (76%), 13R
(3%), and 13S (13%). These figures were in agreement with CP-HPLC anal-
ysis (Fig. 4A). 9S-HPODE can be visualized as formed by oxygen insertion
at C-9 of the 1E,4Z-pentadiene radical with unchanged position of the side
chain at C-9, as discussed below. For comparison, the proportions of prod-
ucts formed upon oxidation of 18:2n−6 by COX-1 were 9R (73%), 9S
(9%), 13R (1%), and 13S (17%) [34].
(9Z,12E)18:2 was oxidized to 13S-HPODE as the main metabolite.
CP-HPLC of the products of (9Z,12E)18:2 is shown in Fig. 4B. Chemical
degradation yielded 4 and 18% of the 9R and 9S stereoisomers, and 5
and 73% of the 13R and 13S stereoisomers. The mechanism of biosynthe-
sis of 13S-HPODE was investigated with [11R-2H] (9Z,12E)18:2. The latter
was oxidized to 13S-HPODE with retention of the 2H label, which also is
the case during oxidation of 18:2n−6. The oxidation at C-13 was there-
fore unlikely due to reverse orientation at the active site of COX-1.
13S-HPODE can be visualized as being formed from (9Z,12E)18:2 by oxy-
gen insertion at C-13 of the 1Z,4Z-pentadiene radical, which is formed
after rotation of the side chain at C-13 (cf. [1]).
3.4. Oxygenation of (9E,12Z)- and (9Z,12E)18:2 and (9E)18:1 by 8R-
DOX
8R-DOX of 5,8-LDS, but not of 7,8-LDS, oxidized (9E,12Z)18:2. The
products were separated by CP-HPLC and analyzed by MS/MS after reduc-
tion to alcohols as shown in Fig. 5. The main metabolite was 9R-HODE
(>80%), but double bond isomers of 8- and 11-HODE were also detected.
We conclude that (9E,12Z)18:2 was mainly subject to hydrogen abstrac-
tion at C-11 and oxygenation at C-9. Such oxygenation at C-9 could not be
detected when 18:2n−6 was the substrate.
As expected from oxidation of 18:1n−9, the 8R-DOX of 7,8- and
5,8-LDS transformed (9Z,12E)18:2 to the 8-hydroperoxy metabolite as
the main product (inset in Fig. 5). (9E)18:1 was oxidized at C-8 by 8R-
DOX of 5,8-LDS but not by 7,8-LDS.
3.5. Oxygenation of (9Z,12E)18:2, [11R-2H] (9Z,12E)18:2, and (9E,12Z)
18:2 by sLOX-1
4 μM sLOX-1 (0.1 M NaBO3; pH 9.0) oxidized 200 μM (9Z,12E)18:2
slowly to the 9-hydroperoxy metabolite with 10E,12E configuration as
the main product (65%) and to 13-HPODE (35%; R/S ratio, 64/36). The
UV absorbance at 235 nm increased linearly at a rate of ~0.03 AU/min.
[11R-2H] (9Z,12E)18:2 was not oxidized at a detectable rate under these
conditions, suggesting inverse substrate orientation in comparison with
18:2n−6 and a prominent D-KIE.
4 μM sLOX-1 oxidized (9E,12Z)18:2 more efficiently than (9Z,12E)
18:2. After a prolonged kinetic lag time (2 min), the UV absorbance
increased linearly at ~0.07 AU/min. CP-HPLC-MS/MS analysis showed
Fig. 1. Oxidation of 18:3n−6, 18:1, and 20:1 by COX-1 monitored by an oxygen elec-
trode. A, 18:1n−9 was not a substrate, as the blank yielded an identical kinetic trace,
whereas 18:1n−6 was oxidized albeit at a slower rate than 18:3n−6. B, 20:1n−6 was
oxidized in analogy with 18:1n−6, but the kinetic trace of 20:1n−9 differed as the
rate increased in the course of 1–2 min.

1512
I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
that 9-HPODE, in an R/S ratio of 60/40, was the main product (95%) along
with the all trans isomer of 13-HPODE (5%). The R/S ratio of 9-HODE is in
agreement with the report by Funk et al., but we could not detect signif-
icant formation of 13-HODE with 9Z,11E configuration [36].
(3 min, 30 °C), which suggested a D-KIE of ~5. Under these conditions,
less than 5% of 18:2n−6 was oxidized.
3.6.4. Oxidation of [8,8-2H2]18:1n−9 by 8R-DOX
The rates of oxidation of [8,8-2H2]18:1n−9 and oleic acid by 8R-DOX
of 7,8- and 5,8-LDS were determined by LC–MS/MS due to unstable read-
ings with the oxygen electrode. The results from these LC–MS/MS analy-
ses supported large D-KIE (>20), as illustrated by the measurements in
Fig. 7B and C. This is in agreement with the D-KIE of 5,8-LDS of A. nidulans
[22].
3.6. D-KIE of COX, 8R-DOX, Mn-LOX, and sLOX-1
3.6.1. Oxidation of [11S-2H]- and [11,11-2H2]18:2n−6 by COX-1
[11S-2H]18:2n−6 was oxidized at a slower rate than 18:2n−6
(Fig. 6A). The D-KIE averaged ~3.3, and this value appeared to be only
slightly larger (~10%) for [11,11-2H2]18:2n−6 as substrate. We conclude
that [11S-2H]18:2n−6 is oxidized with a D-KIE in analogy with
[13,13-2H2]- and [13S-2H]20:4n−6 [16] (Table 1).
3.6.5. Oxidation of [11S-2H]18:2n−6 by 13R-MnLOX and sLOX-1
The relative D-KIE of the two LOX enzymes, measured by formation
of 13-HPODE, differed by less than 10% (Fig. 7D). 13R-MnLOX also forms
15–20% 11S-HPODE lacking UV absorbance at 235 nm. The absolute
D-KIE of this experiment (~53) was in the reported magnitude of the
D-KIE of sLOX-1 [12].
3.6.2. Oxidation of [13,13-2H2]20:1n−6 by COX-1
20:1n−6 was oxidized by COX-1 as described above, but the rate of
oxidation of [13,13-2H2]20:1n−6 was insignificant in comparison with
20:1n−6, as judged by oxygen monitoring (Fig. 6B). Short time incuba-
tion (150 s, 30 °C) and LC–MS/MS analysis confirmed a D-KIE of this
magnitude (~25), as illustrated in Fig. 7A. We conclude that allylic and
bis-allylic carbons are oxidized by COX-1 with different D-KIE (Table 1).
3.7. D-KIE during autoxidation of [13,13-2H2]20:1n−6
[13,13-2H2]20:1n−6 was subject to autoxidation with N-
hydroxyphthalimide, which generates a phthalimido-N-oxyl radical
[37]. Analysis of the molecular anions of all the formed [13,13-2H2]- and
[13-2H]-labeled hydroperoxides (m/z 442 and 443) by LC–MS suggested
a D-KIE of ~9, whereas MS3 analysis of [13-2H]15-HPEME and
3.6.3. Oxidation of [11S-2H]- and [11,11-2H2]18:2n−6 by COX-2
The LC–MS/MS assay showed that [11S-2H]- and [11,11-2H2]
18:2n−6 were both oxidized at an average rate of 19% of 18:2n−6
Fig. 2. The products formed from 18:3n−6 by COX-1 were analyzed by RP-HPLC-MS/MS. A, The top two chromatograms show intensities of selected ions specific for 13-HOTrE and 9-HOTrE,
respectively, during MS/MS analysis (m/z 309→full scan), and the latter was the main metabolite. The bottom chromatogram shows selective ion monitoring of m/z 207 during MS/MS analysis
(m/z 325→full scan). B, MS/MS spectrum of the polar metabolite was identical to the MS/MS spectrum of dinor-PGE1 reported by Murphy and co-workers [35].

I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
1513
Fig. 4. LC–MS analysis of the oxidation of (9E,12Z)18:2 and (9Z,12E)18:2 by COX-1. A,
CP-HPLC-MS/MS with separation of stereoisomers formed from (9E,12Z)18:2. Small
amounts of 9- and 13-[13C18]HODE were added to facilitate identification of stereoisomers
(peak I, 13S-HODE, peak III, 9S-HODE, peak V, 9R-HODE, peak IV, 9-HODE(10E,12E), and
peak II 13-HODE(9E,11E)). B, CP-HPLC-MS/MS with separation of stereoisomers formed
from (9Z,12E)18:2, which were identified by aid of added [13C18]HODE (peak I, 13S-HODE,
peak II, 13R-HODE, and peak III, trans–trans 9-HODE). Separation on Chiralcel OB-H.
4. Discussion
We have investigated the reaction mechanisms of COX-1, COX-2,
8R-DOX, and lipoxygenases with selectively deuterated fatty acids. Oxida-
tion of cis–trans isomers of 18:2n−6 and a selection of mono- and poly-
unsaturated fatty acids was also investigated. The reaction mechanisms of
these enzymes are summarized in Figs. 8–10, and D-KIE in Table 1.
Fig. 3. RP-HPLC-MS/MS analysis of oxidation of 20:1n−6 and 20:1n−9 by COX-1. A,
Analysis of products formed from 20:1n−6. The top chromatogram shows total ion
current (TIC) during MS/MS analysis (m/z 325→full scan), whereas the two bottom
chromatograms show intensities of selected ions specific for 13-HEME and 15-HEME,
respectively. Peak I, 13-HEME, peak II, 15-HEME. B, The chromatogram shows TIC dur-
ing MS/MS analysis (m/z 325→full scan) of products formed from 20:1n−9, and the
main products were identified as marked (13- and 11-HEME).
4.1. Influences of cis–trans isomers of 18:2n−6 and the n−6 double
bond on catalysis
18:2n−6 is oxidized by COX-1 and COX-2 mainly to 9R-HPODE and
to small amounts of 13S-HPODE. (9E,12Z)18:2 was oxidized by COX-1
to 9S-HPODE as the main metabolite. As expected, the 9E double bond
thus changed the configuration of 9-HPODE from 9R to 9S, and this is in
agreement with antarafacial hydrogen abstraction and oxygen insertion.
Unexpectedly, (9Z,12E)18:2 was mainly transformed to the S stereo-
isomer of 13-HPODE and not to the R stereoisomer (Fig. 8). Further
[13,13-2H2]16-HPEME yielded a KIE of ~6. For comparison, the
D-KIE was 5.2–6.8 during autoxidation of [11,11-2H2]18:2 with
the cumylperoxyl radical [38].

1514
I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
8R-DOX of A. fumigatus oxidized (9Z,12E)18:2 in analogy with
18:1n−9, but (9E,12Z)18:2 was transformed to 9R-HPODE as the main
metabolite by 8R-DOX of 5,8-LDS. This occurred apparently by bis-allylic
hydrogen abstraction (Fig. 9), but the relation of hydrogen abstraction
to oxygen insertion was not further investigated. Small amounts of prod-
ucts formed by hydrogen abstraction and oxidation at C-8 and C-11 were
also formed.
COX-1 oxidized 20:1n−6 by hydrogen abstraction at C-13, which was
followed by oxygen insertion at C-13 and, after double bond migration, at
C-15 (Fig. 8); it is noteworthy that the proR hydrogen at C-13 of 20:1n−6
will be designated proS at C-13 of 20:4n−6 due to the Cahn–Ingold–
Prelog nomenclature. These products, 13- and 15-HPEME, were formed
with high stereospecificity (>93% S configuration). 18:1n−6 was oxi-
dized in the same way. The n−6 double bond and chain length seemed
to be important, as 18:1n−9 was not oxidized by COX-1 and 20:1n−9
and 20:3n−9 were oxidized only slowly compared to 20:1n−6 and
20:2n−6.
It is well known that COX-1 accepts chain elongation of 20:4n−6 to
adrenic acid (22:4n−6), which is transformed to dihomo-PGE2 [39].
Chain shortening is also accepted, as COX-1 transformed 18:3n−6 to
dinor-PGE1 although only a small fraction of the peroxyl radical of
9R-HPOTrE(n−6) was converted in this way.
The tertiary structure and catalytic domains of COX-1 and
myeloperoxidase have been described as strikingly similar [40]. It is
therefore of interest that COX-1 abstracts the proR hydrogen of
20:1n−6 (Fig. 8), whereas 8R-DOX abstracts the proS hydrogen of
18:1n−9 (Fig. 9). This observation suggests that 8R-DOX could bind
fatty acids in opposite head-tail orientation than COX-1.
sLOX-1 abstracts the proS hydrogen at C-11 of 18:2n−6 and forms
13S-HPODE as the main metabolite at pH 9 (Fig. 10). At a lower pH
with uncharged carboxyl group, 18:2n−6 binds in the opposite orienta-
tion with biosynthesis of 9S-HPODE [41]. sLOX-1 was reported by Funk et
al. [36] to oxidize the two cis–trans isomers of 18:2n−6 at neutral pH.
To ascertain the importance of the charged carboxyl group we re-
investigated these oxidations at pH 9.
Fig. 5. Oxidation of (9E,12Z)18:2 and (9Z,12E)18:2 by 8R-DOX of 5,8-LDS. A, CP-
HPLC-MS/MS analysis of products formed from (9E,12Z)18:2 after reduction to alco-
hols. The main product was identified as 9R-HODE, whereas MS/MS analysis of peaks
I and II was consistent with 11-HODE and 8-HODE, respectively, presumably with
(9E,12Z) configuration. The inset shows CP-HPLC-MS/MS analysis of products formed
from (9Z,12E)18:2 after reduction to alcohols; the main product was 8-HODE, likely
with 9Z,12E configuration. Analysis on Chiralcel OB-H.
analysis with (9Z,12E)[11R-2H]18:2 showed that this was not due to
reverse orientation in the active site, as the 2H-label was retained
in13S-HPODE. An antarafacial oxygenation mechanism could be exp-
lained, as outlined in Fig. 8, by bond rotation at C-13 with formation of
the 1Z,4Z-pentadiene radical, as previously described in sLOX-1 catalysis
[36]. The effects of bond rotation on radical conformation and the config-
uration of oxygen insertion by COX and LOX were recently reviewed by
Schneider and co-workers [1].
(9Z,12E)18:2 was oxidized slowly to the trans isomer of 9-HODE and
to 13R-HPODE as the two main products (Fig. 10), essentially as reported
previously [36]. [11R-2H] (9Z,12E)18:2 was not oxidized at a detectable
rate, which implies that (9Z,12E)18:2 likely binds to the active site of
sLOX-1 in reverse orientation and presents the proR hydrogen at C-11
for abstraction. 13R-HPODE might be formed by bond rotation at C-13,
as outlined in Fig. 10.
(9E,12Z)18:2 was mainly oxidized at C-9 (>95%) to 9-HPODE in an
R/S ratio of 60/40. 13-HODE was formed with 9E,11E configuration. We
could not find significant biosynthesis of 13-HODE having a 9Z double
bond; this differs from results of bond rotation at C-9/C-10 with forma-
tion of 13-HODE with 9Z configuration observed at neutral pH [36]. The
observation that 9-HPODE was almost racemic suggested that this fatty
acid may partly bind the active site in opposite orientations for hydro-
gen abstraction and oxygenation (Fig. 10).
4.2. D-KIE of lipoxygenases, cyclooxygenases, and 8R-DOX
Labeling of fatty acids with tritium or deuterium has been instrumen-
tal for analysis of reaction mechanisms of fatty acid oxygenases [42]. This
Table 1
Summary of apparent D-KIE of dioxygenase reactions.
Enzymes
Hydrogen abstraction at
Allylic carbons
bis-allylic carbons
COX-1
COX-2
8R-DOX
sLOX-1
MnLOX
~3
~5
ND
~53
~53
~25
ND
>20
–
Fig. 6. Oxidation of deuterated fatty acids by COX-1 with monitoring of oxygen con-
sumption. A, Oxidation of 18:2n−6 and [11S-2H]18:2n−6. B, Oxidation of 20:1n−6
and [13,13-2H2]20:1n−6.
–
ND, not determined. –, oxidized at insignificant rates compared to bis-allylic carbons.

I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
1515
Fig. 7. LC–MS/MS analysis of apparent D-KIE of COX-1 and 8R-DOX and UV analysis of relative D-KIE of 13R-MnLOX and sLOX-1. A, LC–MS analysis of the oxidation of 20:1n-−6 and
[13,13-2H2]20:1n−6 by COX-1 with aid of an internal standard (13-HODE). The first and second chromatograms show the signal intensities of the internal standard and 13- and
15-HEME. The third and fourth chromatograms show the signal intensities of the internal standard and deuterated 13- and 15-HEME. The signal intensities suggest a D-KIE of ~25,
and the illustrated chromatograms are median values of triplicate analysis (30 °C, 150 s). B, Oxidation of 18:1n−9 and [8,8-2H2]18:1n−9 by 8R-DOX of 7,8-LDS (mean of triplicate
analysis; +4 °C) at different time points assessed with an internal standard (13-HODE). C, Oxidation of 18:1n−9 and [8,8-2H2]18:1n−9 by 8R-DOX of 5,8-LDS. D, UV analysis of
products formed from 18:2n−6 and [11S-2H]18:2n−6 by 13R-MnLOX and sLOX-1. The latter was incubated with double enzyme concentration.
led to the discovery of the prominent D-KIE of sLOX-1 [12,43]. The D-KIE
of 13R-MnLOX and sLOX-1 appeared to be rather similar (Fig. 7D). We
conclude that the two different catalytic metals may not alter the D-KIE
under our experimental conditions (23 °C, pH 9.0) although the redox
potentials of the free metals, Mn3+/Mn2+ and Fe3+/Fe2+, differ by a
factor of 2 [24].
In analogy with the allylic oxidation of 20:1n−6 by COX-1, 8R-DOX of
7,8- and 5,8-LDS oxidized the allylic C-8 of 18:1n−9 with a large D-KIE.
This is also in agreement with a recent report on 5,8-LDS of A. nidulans1
[22]. The D-KIE of COX during oxidation of an allylic position can be com-
pared to the D-KIE during oxidation at C-2 of [2,2-2H2]16:0 by α-DOX,
which also requires more energy than oxidation of bis-allylic carbons.
D-KIE of α-DOX averaged 31 at pH 7.2 [17]. The large apparent D-KIE of
COX-1, 8R-DOX, and α-DOX of at least 20 suggest a significant contribu-
tion of hydrogen tunneling to these reactions. The midpoint redox poten-
tial of Tyr•/Tyr is 0.94 V (pH 7.0; 25 °C) [44], which is apparently
sufficient for oxidation of both allylic and bis-allylic carbons, whereas
the redox potential of sLOX-1 has been estimated to be ~0.6 V [45]. In
addition, hydrogen donor and acceptor distances and the reversibility of
hydrogen transfer may conceivably differ between COX and LOX.
We found that COX-1 oxidized the allylic C-13 of 20:1n−6 with at
least a 7-fold larger D-KIE than the bis-allylic C-11 of 18:2n−6 under
identical experimental conditions. These results suggest that D-KIE differ
between allylic and bis-allylic hydrogen abstraction. The contribution of
tunneling to D-KIE could be influenced by the energy necessary for hydro-
gen abstraction, the surface potential of the enzyme, and by kinetic
factors. Kulmacz and co-workers investigated the oxidation of [13S-2H]-
and [13,13-2H2]20:4n−6 by purified COX-1 at different experimental
conditions, and reported a weakly temperature-dependent D-KIE of 2–3
[16]. In contrast, it was shown that D-KIE of recombinant COX-2 with
[2H31]18:2n−6 as substrate varied with the experimental conditions [18],
but that it was larger than reported for 20:4n−6 [16]. These results may
appear contradictory, but the perdeuterated [2H31]18:2n−6 may con-
ceivably interact with the surface of the active site of COX in a less bene-
ficial manner for catalysis than [13S-2H]20:4n−6, [11S-2H]18:2n−6,
and [11,11-2H2]18:2n−6. We found that the latter two were oxidized
by COX-1 and COX-2 with apparent D-KIE of 3–5. Remote deuterium
atoms may exert secondary effects, as [5,5-2H2]18:1n−9 was oxidized
at C-8 by 5,8-LDS of A. nidulans1at a lower rate than 18:1n−9 [22].
5. Conclusion
COX-1 binds mono- and polyunsaturated fatty acids with the omega
end first. The cis n−6 double bond facilitated allylic oxidation by COX-1
in comparison with cis n−9 fatty acids. bis-Allylic oxidation by COX-1
was catalyzed more efficiently and with insignificant hydrogen tunneling
compared to allylic oxidation. COX-1 oxidized (9Z,12E)18:2 with bond ro-
tation at C-13, whereas 8R-DOX shifted hydrogen abstraction of (9E,12Z)
18:2 to C-11 and oxygen insertion to C-9. Fe or Mn as catalytic metal of
LOX appeared to have little influence on D-KIE.

1516
I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
Fig. 8. Overview of hydrogen abstraction and formation of major oxidation products of
20:1n−6, (9E,12Z)18:2 and (9Z,12E)18:2 by COX-1. A, The proR hydrogen of C-13 is
abstracted and oxygen is inserted at C-13 and at C-15. B, Oxidation of (9E,12Z)-18:2
to 9S-HPODE as the main metabolite. C, Oxidation of (9Z,12E)-18:2 to 13S-HPODE as
the main metabolite. [11R-2H] (9Z,12E)-18:2 was oxidized with retention of the deute-
rium label, supporting unchanged orientation in the active site. 13S-HPODE could be
formed by bond rotation at C-13 of the pentadiene radical as indicated.
Fig. 10. Overview of hydrogen abstraction and formation of major oxidation products
of 18:2n−6, (9E,12Z)18:2 and (9Z,12E)18:2 by sLOX-1. A, 18:2n−6 is oxidized at
C-13 at alkaline pH (top) in normal orientation and at C-9 at acidic pH (bottom) to
9S-HPODE by inverse orientation in the active site. B and C, The two cis–trans fatty
acids are mainly oxidized at the trans double bonds. (9Z,12E)18:2 in B appeared to
be oxidized in both orientations as this fatty acid was oxidized to 9-HPODE in an R/S
ratio of 60/40. (9E,12Z)18:2 in C was likely oxidized by abstraction of the proS hydro-
gen at C-11 in the reverse orientation, as [11R-2H](9E,12Z)18:2 was a poor substrate.
13R-HPODE might be formed by bond rotation at C-13 as indicated.
Acknowledgements
This work was supported by VR (03X-06523) and by the Knut and
Alice Wallenberg Foundation (KAW 2004.0123).
References
[1] C. Schneider, D.A. Pratt, N.A. Porter, A.R. Brash, Control of oxygenation in lipoxygenase
and cyclooxygenase catalysis, Chem. Biol. 14 (2007) 473–488.
[2] A. Andreou, I. Feussner, Lipoxygenases — structure and reaction mechanism, Phy-
tochemistry 70 (2009) 1504–1510.
[3] J.Z. Haeggström, C.D. Funk, Lipoxygenase and leukotriene pathways: biochemis-
try, biology, and roles in disease, Chem. Rev. 111 (2011) 5866–5898.
[4] W.L. Smith, Y. Urade, P.J. Jakobsson, Enzymes of the cyclooxygenase pathways of
prostanoid biosynthesis, Chem. Rev. 111 (2011) 5821–5865.
Fig. 9. Overview of formation of major oxidation products of 18:1n−9 and (9E,12Z)
18:2 by 8R-DOX. A, Hydrogen abstraction and oxygenation at C-8 of 18:1n−9 occur
in an antarafacial way [21]. B, (9E,12Z)18:2 is oxidized by hydrogen abstraction at
C-11 and oxygen insertion at C-9 with formation of 9R-HPODE; the steric relation of
hydrogen abstraction and oxygen insertion has not been determined.

I. Hoffmann et al. / Biochimica et Biophysica Acta 1821 (2012) 1508–1517
1517
[5] U. Garscha, F. Jernerén, D. Chung, N.P. Keller, M. Hamberg, E.H. Oliw, Identification
of dioxygenases required for Aspergillus development. studies of products, stereo-
chemistry, and the reaction mechanism, J. Biol. Chem. 282 (2007) 34707–34718.
[6] F. Brodhun, I. Feussner, Oxylipins in fungi, FEBS J. 278 (2011) 1047–1063.
[7] J. Vicente, T. Cascon, B. Vicedo, P. Garcia-Agustin, M. Hamberg, C. Castresana, Role
of 9-lipoxygenase and alpha-dioxygenase oxylipin pathways as modulators of
local and systemic defense, Mol. Plant 5 (2012) 914–928.
[26] M. Hamberg, Stereochemistry of hydrogen removal during oxygenation of
linoleic acid by singlet oxygen and synthesis of 11(S)-deuterium-labeled linoleic
acid, Lipids 46 (2011) 201–206.
[27] E.H. Oliw, L. Hörnsten, H. Sprecher, Oxygenation of 5,8,11-eicosatrienoic acid by prosta-
glandin H synthase-2 of ovine placental cotyledons: isolation of 13-hydroxy-5,8,11-
eicosatrienoic and 11-hydroxy-5,8,12-eicosatrienoic acids, J. Chromatogr. B Analyt.
Technol. Biomed. Life Sci. 690 (1997) 332–337.
[8] M. Hamberg, I. Ponce de Leon, M.J. Rodriguez, C. Castresana, Alpha-dioxygenases,
Biochem. Biophys. Res. Commun. 338 (2005) 169–174.
[28] T. Caruso, A. Spinella, Cs2CO3 promoted coupling reactions for the preparation of
skipped diynes, Tetrahedron 59 (2003) 7787–7790.
[9] I. Ivanov, D. Heydeck, K. Hofheinz, J. Roffeis, V.B. O'Donnell, H. Kuhn, M. Walther, Molec-
ular enzymology of lipoxygenases, Arch. Biochem. Biophys. 503 (2010) 161–174.
[10] C. Su, E.H. Oliw, Manganese lipoxygenase. Purification and characterization, J. Biol.
Chem. 273 (1998) 13072–13079.
[29] C.A. Brown, V.K. Ahuja, Catalytic-hydrogenation. 6. Reaction of sodium-borohydride
with nickel salts in ethanol solution — P-2 nickel, a highly convenient, new, selective
hydrogenation catalyst with great sensitivity to substrate structure, J. Org. Chem. 38
(1973) 2226–2230.
[11] A. Wennman, F. Jernerén, M. Hamberg, E.H. Oliw, Catalytic convergence of Mn-
and iron lipoxygenases by replacement of a single amino acid, J. Biol. Chem. 287
(2012) 31757–31765.
[30] E.J. Corey, G. Schmidt, Useful procedures for the oxidation of alcohols involving
pyridinium dichromate in aprotic media, Tetrahedron Lett. (1979) 399–402.
[31] M. Hamberg, Stereochemistry of oxygenation of linoleic acid catalyzed by
[12] M.J. Knapp, J.P. Klinman, Environmentally coupled hydrogen tunneling. Linking
catalysis to dynamics, Eur. J. Biochem. 269 (2002) 3113–3121.
prostaglandin-endoperoxide H synthase-2, Arch. Biochem. Biophys. 349
(1998) 376–380.
[13] F.H. Westheimer, The magnitude of the primary kinetic isotope effect for com-
pounds of hydrogen and deuterium, Chem. Rev. 61 (1961) 265–273.
[14] M.P. Meyer, D.R. Tomchick, J.P. Klinman, Enzyme structure and dynamics affect hydro-
gen tunneling: the impact of a remote side chain (I553) in soybean lipoxygenase-1,
Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 1146–1151.
[15] C. Su, M. Sahlin, E.H. Oliw, A protein radical and ferryl intermediates are generat-
ed by linoleate diol synthase, a ferric hemeprotein with dioxygenase and hydro-
peroxide isomerase activities, J. Biol. Chem. 273 (1998) 20744–20751.
[16] G. Wu, J.M. Lu, W.A. van der Donk, R.J. Kulmacz, A.L. Tsai, Cyclooxygenase reaction
mechanism of prostaglandin H synthase from deuterium kinetic isotope effects,
J. Inorg. Biochem. 105 (2011) 382–390.
[32] M. Hamberg, Steric analysis of hydroperoxides formed by lipoxygenase oxygena-
tion of linoleic acid, Anal. Biochem. 43 (1971) 515–526.
[33] M.G. Malkowski, E.D. Thuresson, K.M. Lakkides, C.J. Rieke, R. Micielli, W.L. Smith, R.M.
Garavito, Structure of eicosapentaenoic and linoleic acids in the cyclooxygenase site
of prostaglandin endoperoxide H synthase-1, J. Biol. Chem. 276 (2001) 37547–37555.
[34] M. Hamberg, B. Samuelsson, Stereochemistry in the formation of 9-hydroxy-
10,12-octadecadienoic acid and 13-hydroxy-9,11-octadecadienoic acid from linoleic
acid by fatty acid cyclooxygenase, Biochim. Biophys. Acta 617 (1980) 545–547.
[35] R.C. Murphy, R.M. Barkley, K. Zemski Berry, J. Hankin, K. Harrison, C. Johnson, J. Krank, A.
McAnoy, C. Uhlson, S. Zarini, Electrospray ionization and tandem mass spectrometry of
eicosanoids, Anal. Biochem. 346 (2005) 1–42.
[17] A. Gupta, A. Mukherjee, K. Matsui, J.P. Roth, Evidence for protein radical-mediated
nuclear tunneling in fatty acid alpha-oxygenase, J. Am. Chem. Soc. 130 (2008)
11274–11275.
[36] M.O. Funk Jr., J.C. Andre, T. Otsuki, Oxygenation of trans polyunsaturated fatty
acids by lipoxygenase reveals steric features of the catalytic mechanism, Bio-
chemistry 26 (1987) 6880–6884.
[18] H.H. Danish, I.S. Doncheva, J.P. Roth, Hydrogen tunneling steps in cyclooxygenase-2
catalysis, J. Am. Chem. Soc. 133 (2011) 15846–15849.
[19] F. Brodhun, C. Göbel, E. Hornung, I. Feussner, Identification of PpoA from Aspergil-
lus nidulans as a fusion protein of a fatty acid heme dioxygenase/peroxidase and a
cytochrome P450, J. Biol. Chem. 284 (2009) 11792–11805.
[20] I. Hoffmann, F. Jernerén, U. Garscha, E.H. Oliw, Expression of 5,8-LDS of Aspergillus
fumigatus and its dioxygenase domain. A comparison with 7,8-LDS, 10-dioxygenase,
and cyclooxygenase, Arch. Biochem. Biophys. 506 (2011) 216–222.
[21] M. Hamberg, L.Y. Zhang, I.D. Brodowsky, E.H. Oliw, Sequential oxygenation of linoleic
acid in the fungus Gaeumannomyces graminis: stereochemistry of dioxygenase and
hydroperoxide isomerase reactions, Arch. Biochem. Biophys. 309 (1994) 77–80.
[22] A. Nadler, C. Koch, F. Brodhun, J.D. Wehland, K. Tittmann, I. Feussner, U.
Diederichsen, Influence of substrate dideuteration on the reaction of the bifunc-
tional heme enzyme psi factor producing oxygenase A (PpoA), Chembiochem
12 (2011) 728–737.
[23] D.A. Pratt, J.H. Mills, N.A. Porter, Theoretical calculations of carbon–oxygen bond
dissociation enthalpies of peroxyl radicals formed in the autoxidation of lipids,
J. Am. Chem. Soc. 125 (2003) 5801–5810.
[24] E.H. Oliw, F. Jernerén, I. Hoffmann, M. Sahlin, U. Garscha, Manganese lipoxygenase
oxidizes bis-allylic hydroperoxides and octadecenoic acids by different mechanisms,
Biochim. Biophys. Acta 1811 (2011) 138–147.
[25] E.H. Oliw, A. Wennman, I. Hoffmann, U. Garscha, M. Hamberg, F. Jernerén,
Stereoselective oxidation of regioisomeric octadecenoic acids by fatty acid
dioxygenases, J. Lipid Res. 52 (2011) 1995–2004.
[37] F. Recupero, C. Punta, Free radical functionalization of organic compounds cata-
lyzed by N-hydroxyphthalimide, Chem. Rev. 107 (2007) 3800–3842.
[38] H. Kitaguchi, K. Ohkubo, S. Ogo, S. Fukuzumi, Additivity rule holds in the hydro-
gen transfer reactivity of unsaturated fatty acids with a peroxyl radical: mecha-
nistic insight into lipoxygenase, Chem. Commun. (Camb) (2006) 979–981.
[39] W.B. Campbell, J.R. Falck, J.R. Okita, A.R. Johnson, K.S. Callahan, Synthesis of
dihomoprostaglandins from adrenic acid (7,10,13,16-docosatetraenoic acid) by
human endothelial cells, Biochim. Biophys. Acta 837 (1985) 67–76.
[40] D. Picot, P.J. Loll, R.M. Garavito, The X-ray crystal structure of the membrane pro-
tein prostaglandin H2 synthase-1, Nature 367 (1994) 243–249.
[41] H.W. Gardner, Soybean lipoxygenase-1 enzymically forms both (9S)- and (13S)-hydro-
peroxides from linoleic acid by a pH-dependent mechanism, Biochim. Biophys. Acta
1001 (1989) 274–281.
[42] A.R. Brash, C. Schneider, M. Hamberg, Applications of stereospecifically-labeled
fatty acids in oxygenase and desaturase biochemistry, Lipids 47 (2012)
101–116.
[43] C.C. Hwang, C.B. Grissom, Unusually large deuterium-isotope effect in soybean
lipoxygenase is not caused by a magnetic isotope effect, J. Am. Chem. Soc. 116
(1994) 795–796.
[44] M.R. Defelippis, C.P. Murthy, M. Faraggi, M.H. Klapper, Pulse radiolytic measurement of
redox potentials - the tyrosine and tryptophan radicals, Biochemistry 28 (1989)
4847–4853.
[45] M.J. Nelson, Catecholate complexes of ferric soybean lipoxygenase 1, Biochemis-
try 27 (1988) 4273–4278.