Cryptoregiochemistry of a Brassica napus fatty acid desaturase (FAD3): A kinetic isotope effect study

Article abstract of DOI:10.1039/b008111f

α-Linolenic acid ((Z,Z,Z)-octadeca-9,12,15-trienoic acid) is biosynthesized by a series of regio- and stereoselective dehydrogenation reactions which are catalyzed by a set of enzymes known as fatty acid desaturases. As part of ongoing research into the mechanism of these remarkable catalysts, we have examined the cryptoregiochemistry (site of initial oxidation) of extraplastidial ω - 3 desaturation as it occurs in the commercially important plant Brassica napus (oilseed rape or canola). The individual deuterium kinetic isotope effects associated with the C-H bond cleavages at C-15 and C-16 of a thiaoleoyl analogue were measured using a convenient in vivo yeast expression system. Competition experiments using appropriately deuterium-labelled 7-thia substrates revealed a large kinetic isotope effect (KIE) (k_H/k_D = 7.5 ± 0.4) for the C-H bond-breaking step at C-15 while the C-H bond cleavage at C-16 was found to be insensitive to deuterium substitution (k_H/k_D = 1.0 ± 0.14). These results point to C-15 as the site of initial oxidation in ω - 3 desaturation since the first chemical step in this type of reaction is rupture of a strong, unactivated C-H bond - an energetically difficult process which typically exhibits a large KIE.

Full text of DOI:10.1039/b008111f

View Article Online / Journal Homepage / Table of Contents for this issue
Cryptoregiochemistry of a Brassica napus fatty acid desaturase
(FAD3): a kinetic isotope effect study
Christopher K. Savile,a Darwin W. Reed,b Dauenpen Meesapyodsuk,b Patrick S. Covello*b and
Peter H. Buist*a
1
a Ottawa-Carleton Chemistry Institute, Department of Chemistry, Carleton University,
1125 Colonel By Drive, Ottawa, Ontario, Canada K1S 5B6
b National Research Council of Canada, Plant Biotechnology Institute, 110 Gymnasium Place,
Saskatoon, Saskatchewan, Canada S7N 0W9
Received (in Cambridge, UK) 9th October 2000, Accepted 13th March 2001
First published as an Advance Article on the web 17th April 2001
α-Linolenic acid ((Z,Z,Z)-octadeca-9,12,15-trienoic acid) is biosynthesized by a series of regio- and stereoselective
dehydrogenation reactions which are catalyzed by a set of enzymes known as fatty acid desaturases. As part of
ongoing research into the mechanism of these remarkable catalysts, we have examined the cryptoregiochemistry (site
of initial oxidation) of extraplastidial ω Ϫ 3 desaturation as it occurs in the commercially important plant Brassica
napus (oilseed rape or canola). The individual deuterium kinetic isotope effects associated with the C–H bond
cleavages at C-15 and C-16 of a thiaoleoyl analogue were measured using a convenient in vivo yeast expression
system. Competition experiments using appropriately deuterium-labelled 7-thia substrates revealed a large kinetic
isotope effect (KIE) (kH/kD = 7.5 0.4) for the C–H bond-breaking step at C-15 while the C–H bond cleavage at C-16
was found to be insensitive to deuterium substitution (kH/kD = 1.0 0.14). These results point to C-15 as the site of
initial oxidation in ω Ϫ 3 desaturation since the first chemical step in this type of reaction is rupture of a strong,
unactivated C–H bond—an energetically difficult process which typically exhibits a large KIE.
Introduction
The search for biocatalysts which can activate C–H bonds in a
selective manner remains an important scientific objective.1 The
use of soluble and membrane-bound forms of both the cyto-
chromes P-4502 and the emerging non-heme diiron family of
enzymes3 appears promising in this regard. A particularly
impressive subset of the latter class includes the fatty acid
desaturases which are capable of selectively dehydrogenating
the non-activated positions of conformationally mobile sub-
strates.4 Some biotechnologically important examples of
desaturase-mediated chemistry can be found in the biosynthetic
pathway of the common plant fatty acids as shown in Scheme 1.
The latter two reactions in this scheme occur in parallel on
different lipid substrates within and outside the chloroplast or
plastid. It is clear that the desaturation reaction is executed with
exquisite chemo-, regio- and stereoselectivity. Importantly,
these highly desirable properties are potentially tunable by pro-
tein engineering and recently, the first steps in this direction
have been taken.5a,b Critical to the design and interpretation of
such experiments is the identification of the methylene group
Scheme 1
which is attacked first during the introduction of the double
bond. Three complementary methods for determining this par-
ameter have been developed by the Buist group,6a–c based on the
C–H activation step and is completed by rapid collapse of a
concept that desaturation represents a mechanistic variation6d
short-lived radical “intermediate” or its organoiron equivalent
(not shown) through loss of a second hydrogen to give an ole-
of the more well-known hydroxylation reaction (Scheme 2). Our
fin.§ Hydroxylation would proceed by hydroxy rebound (SH2)
model features a highly reactive iron oxo species † which is
asymmetrically located with respect to two methylene groups
forced into a gauche‡ conformation by the active site contours of
to give alcohol. Very recently,9a support for the intermediacy
of a radical intermediate in the latter process as it occurs in
the membrane-bound, non-heme diiron-containing ω-alkane
the enzyme. Oxidation is initiated by an energetically difficult,
hydroxylase9b has been obtained. According to our mechanistic
scheme, the first step in desaturation reactions should be much
more affected by isotopic substitution than the second as might
† We have chosen to use a generic iron oxo representation of the active
oxidant although other structures are possible.7
‡ Occasionally, as in the case of insect pheromones, (E)-olefins are
formed which presumably requires that the substrate adopt an anti con-
formation.8 In all cases examined to date, desaturation takes place via
syn-removal of adjacent hydrogens.8
§ A full discussion of other possible mechanisms for desaturation
is found in ref. 6d. The precise nature of the switch controlling the
olefin : alcohol product ratio remains a matter of speculation.
1116
J. Chem. Soc., Perkin Trans. 1, 2001, 1116–1121
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b008111f

View Article Online
this enzyme using the KIE approach. In this paper, we docu-
ment the design, synthesis and desaturation of our crypto-
regiochemical probes.¶
Results and discussion
Previous studies6b,7,11f have shown that baker’s yeast, Sac-
charomyces cerevisiae, is an extremely convenient in vivo
microbial system for studying the kinetic isotope effects associ-
ated with fatty acid desaturation. There are two principal
reasons for this: firstly, cultures of this organism readily
incorporate labelled exogenous fatty acid methyl esters and
convert them to the appropriate substrate required by the
desaturase under study. Secondly, S. cerevisiae has proven to be
the eukaryotic host of choice for the heterologous expression of
desaturases from a diverse array of sources.11f,14,15 Our KIE
method involves incubating an equimolar mixture of the
appropriate regiospecifically dideuterated (–CD2–) fatty acid
analogue and the non-deuterated parent compound with cul-
tures of a yeast strain bearing the appropriate desaturase gene.
Evaluation of the d1/d0 ratio in the olefinic products by mass
spectral examination allows one to compare the competitive
deuterium kinetic isotope effect for each individual C–H
cleavage involved in desaturation. We typically use deuterated
substrates in which a remote methylene group is replaced by a
sulfur atom in order to eliminate interference by endogenous
olefinic fatty acids in the mass spectral analyses. (It is known
that substitution of a C-7 methylene group with a sulfur atom
in the substrate does not change the cryptoregiochemistry of
double bond introduction11d,e). In general, our estimates
of the KIE correlate well with those originally measured by
Strittmatter using 9,9,10,10-d4-stearoyl CoA and a highly puri-
fied, reconstituted rat liver ∆9 desaturase system.11i In vitro
measurements using a related sterol desaturase has revealed
that the primary deuterium kinetic isotope effects on V/K and
V are similar.11g
A preliminary experiment was carried out involving the
administration of methyl 7-thiastearate 216 (100 mg LϪ1) to cul-
tures (50 cm3) of the pRS131/MKP-o strain of S. cerevisiae15
incubated at 20 ЊC for 3 days for relatively rapid growth and
then at 15 ЊC for a further 3 days to reach saturation at a tem-
perature which has been found to give better substrate conver-
sion rates. The cell cultures were harvested by centrifugation
and the cellular fatty acids were isolated via a standard
hydrolysis–methylation sequence.15 Analysis of the fatty acids
as methyl esters by GC-MS revealed a typical fatty acid profile
for this organism (Fig. 1A) when it is grown in the presence of
thia fatty acid analogues. We were pleased to observe that the
7-thiastearoyl starting material 2 had been transformed into the
corresponding 7-thiaoleate 3 by the endogenous yeast ∆9
desaturase, as expected,16 and then desaturated further to a
9,15-dienoic material 4 by FAD3 (Scheme 3) as determined by
mass spectral analysis.11f ¶ The amount of 4 could be signifi-
cantly enriched by HPLC fractionation of the sample (Fig. 1B).
Control experiments with the yeast strain MKP-o containing
the pSE936 plasmid without the fad3 gene indicated that the
production of 4 was dependent on the presence of the FAD3
enzyme.
Scheme 2
be expected, a priori, for a disproportionation reaction.10
Indeed, precisely such a pattern of kinetic isotope effects (one
large, one negligible) has been found by ourselves and others for
all membrane-bound desaturases examined to date6b,6d,11a–g and
this information has been used to pinpoint the site of initial
oxidation. In a number of cases, corroborating evidence is
available to support the cryptoregiochemical assignment: that
is, when a desaturase behaves as an oxygenase or vice versa, the
regioselectivity of oxygenation matches that of the initial
hydrogen abstraction step involved in desaturation.5b,6a,6d,11e,11h
In addition, a strong primary deuterium KIE (kH/kD = 7.8) on
the initial hydrogen abstraction step has been found in the case
of the alkane ω-hydroxylase (vide supra).9b Finally, our observ-
ation of a strong α-fluoro effect6c at the predicted site of initial
oxidation for a ∆9 desaturase is indicative of hydrogen abstrac-
tion by a strongly electrophilic reagent such as the putative iron
oxo species.
An interesting trend has emerged from these studies: in all
cases examined to date, it appears that it is the C–H bond clos-
est to the acyl terminus which is attacked first by the iron oxi-
dant. Most of the desaturases which have been analyzed are of
the ∆x type; i.e. the double bond is introduced at a fixed distance
from the carbonyl recognition point over a range of substrate
chain lengths (e.g. ∆9, Scheme 1). Recently, we have looked at
the cryptoregiochemistry of a fatty acid desaturase from the
nematode Caenorhabditis C. elegans (FAT-1) which inserts a
double bond using the terminal methyl group as a reference
point (ω Ϫ x type). The initial site of oxidation in this case was
again found to be the carbon closest to C-1 as determined by
our KIE methodology.11f This is an intriguing result since it
suggests that two disparate subclasses of desaturases might
share common cryptoregiochemistry. To test the generality of
this phenomenon, we thought it would be useful to study the
extraplastidial ω Ϫ 3 desaturase (FAD3) from plants—an
enzyme which is involved in the biosynthesis of α-linolenic acid
(ALA, (Z,Z,Z)-octadeca-9,12,15-trienoic acid) 1 (Scheme 1) in
the endoplasmic reticulum. ALA plays an important role in
plant12 and animal13 physiology and has found many appli-
cations in the coatings industry. Comparison of predicted
amino acid sequences14 suggests that the plant and nematode
ω Ϫ 3 desaturases are two structurally divergent proteins, which
were independently derived from ∆12 desaturases. Furthermore,
the enzymes are thought to act on different lipid substrates—
the plant extraplastidial enzyme utilizing phosphatidylcholine
derivatives and the animal enzyme recognizing acyl-coenzyme
A thioesters. Consequently, it is of interest to probe the
cryptoregiochemistry of the plant desaturase and compare it to
a corresponding enzyme from the animal kingdom. Covello et
al.15 have recently provided a convenient expression system for
the FAD3 desaturase from Brassica napus, thus offering a
unique opportunity to determine the cryptoregiochemistry of
The two regiospecifically dideuterated methyl 7-thiastearates
([15,15-2H2]-2, [16,16-2H2]-2) required for the KIE study were
prepared via synthetic routes similar to that previously
described6d in overall yields of 4% and 3% respectively
(Scheme 4).
¶ This work was presented at the recent 51st Harden Conference on
Fatty Acid Desaturases held in Wye, Kent, UK, 30 July to 2 August,
2000.
|| S. cerevisiae does not possess a native ∆12 desaturase. It has been
shown11f,15 that the ω Ϫ 3 desaturase will operate on substrates which
lack a (Z)-double bond at C-12,13.
J. Chem. Soc., Perkin Trans. 1, 2001, 1116–1121
1117

View Article Online
Fig. 1 (A) GC-MS analysis of FAME (fatty acid methyl esters), total
extract derived from pRS131/MKP-o yeast cultures supplied with
undeuterated methyl 7-thiastearate 2. The thia starting material,
monoenoic intermediate and the dienoic product are indicated as S18:0,
S18:1 and S18:2, respectively. 16:0 = methyl palmitate (hexadecanoate);
16:1 = methyl palmitoleate ((Z)-hexadec-9-enoate); 18:0 = methyl
stearate (octadecanoate); 18:1 = methyl oleate ((Z)-octadec-9-enoate).
S18:1 appeared as an unresolved shoulder on the peak due to S18:0.
(B) GC-MS analysis of HPLC fraction (reaction time = 4–5 min)
containing S18:2. (C) GC-MS analysis of HPLC fraction (reaction
time = 5–6 min) containing S18:1; See Experimental for culture
conditions and instrumental parameters for GC-MS and HPLC.
Scheme 4 Reagents and conditions: (a) (i) LiAlD4; (ii) HBr, H2SO4;
(iii) 8-bromooct-1-ene, Mg, THF; (iv) KMnO4, H+; (v) LiAlH4; (vi)
TsCl, pyridine; (vii) n-BuLi (2 equiv.), THF; (viii) BF3–MeOH. (b) (i)
PBr3; (ii) ethylmagnesium bromide, THF; (iii) KMnO4, H+; (iv) LiAlH4;
(v) TsCl, pyridine; (vi) n-BuLi (2 equiv.), THF; (vii) BF3–MeOH.
A ∼1 : 1 mixture of each deuterated material with its non-
deuterated parent (5 mg) was administered to growing cultures
(50 cm3) of the S. cerevisiae transformant (pRS131/MKP-o)
using conditions identical to that of the trial experiment. All
incubations were carried out three times. The deuterium con-
tent of the olefinic fatty acid methyl esters in the cellular lipid
extract was assessed by GC-MS as described in the Experi-
mental section. As expected, the d2 : d0 ratio of the methyl
7-thiaoleate isotopomers (3, S18:1) was essentially identical
to that of the starting material in both incubations, indi-
cating that the condition for low conversion (<10%) which is
required for these types of competitive KIE measurements has
been met.19 Indeed, as can be seen in Fig. 1, the amount of
ω Ϫ 3-desaturated product (4, S18:2) is less than 1% of the
thiaoleate (3, S18:1) fraction and thus no significant isotopic
enrichment is observed for the latter compound (vide infra).
Mass spectral analysis of the product of the second desatur-
ation reaction (4, S18:2) revealed that in both incubations this
material consisted entirely of a d0–d1 mixture as expected,
Scheme 3
The key step in these syntheses is the Grignard coupling reac-
tion between two alkyl bromides which depends on the presence
of a Cu() catalyst for success.17 The final products were purified
by careful flash chromatography to remove small amounts of
C-17 material produced by overoxidation in the KMnO4-
mediated alkene cleavage step.18 GC-MS analysis of the final
deuterated products revealed that each isotopomer consisted
essentially entirely of dideuterated species.
1118
J. Chem. Soc., Perkin Trans. 1, 2001, 1116–1121

View Article Online
Table 1 Intermolecular kinetic isotope effects on the ω Ϫ 3 desatur-
ation of 7-thiaoctadec-9-enoate catalyzed by B. napus FAD3a
running spectra of mixtures (1 : 2) of labelled and unlabelled
material. Mass spectra of substrates were obtained by GC-MS
using a Kratos Concept 1H mass spectrometer equipped with a
70 eV EI ionization chamber. Samples were introduced using an
HP 5980 Series 2 gas chromatograph equipped with a J. & W.
30 m × 0.21 mm, DB-5 capillary column. GC-MS analyses of
fatty acid methyl esters (FAME) from incubation experiments
were performed using a Fisons VG TRIO 2000 mass spec-
trometer (VG Analytical UK) controlled by Masslynx version
2.0 software, coupled to a GC 8000 Series gas chromatograph
equipped with a 30 m × 0.253 mm DB-23 (0.25 µm film thick-
ness) fused silica column (temperature program isothermal
200 ЊC for 1 min, gradient 10 ЊC minϪ1 to 240 ЊC and then
isothermal at 240 ЊC for 25 min). Deuterium content was
measured using selected ion monitoring in EI mode (70 eV) of
pertinent ion clusters with a cycle time of 0.1 s per channel
(4 channels for methyl 7-thiaoleate and 5 channels for methyl
7-thiaisolinoleate) corresponding to ca. 20–25 scans per peak.
The integrated intensities of the individual ions were corrected
for natural isotopic abundance and the isotopic ratios were
determined using these corrected intensities for the ions: m/z
FAME mixture supplied
Isotope effect
15,15-d2S18:0Me–S18:0Me
16,16-d2S18:0Me–S18:0Me
7.5 0.4
1.0 0.14
a Assays and intermolecular isotope effect calculations were performed
as described under Experimental. Values are the means and standard
deviations of separate calculations for 3 cultures.
indicating a loss of one deuterium from the d2-substrate.
Product kinetic isotope effects (kH/kD) were calculated using the
ratio: [% d0 (product)/% d1 (product)] : [% d0 (substrate)/% d2
(substrate)] and this analysis (Table 1) indicated the presence of
a large deuterium isotope effect (7.5 0.4, average of 3 experi-
ments)** for the carbon–hydrogen bond cleavage at C-15 while
the C16–H bond breaking step was shown to be essentially
insensitive to deuterium substitution (KIE = 1.0 0.14, average
of 3 experiments). According to our mechanism (Scheme 2),
these results demonstrate that the site of initial oxidation for
ω Ϫ 3 desaturation is C-15 rather than C-16. This result echoes
a similar finding obtained recently for an ω Ϫ 3 desaturase from
the animal kingdom where the site of initial oxidation for the
C. elegans desaturase was also determined to be at C-15 for a
C18 substrate (kH/kD = 7.8 0.4 (C-15), kH/kD = 0.99 0.04
(C-16)). The common cryptoregiochemistry shared between
these two enzymes points to similar active site motifs despite
the differences in substrate acyl head group recognition (see
Introduction). The precise details of these architectural elem-
ents will hopefully be unveiled when the 3-D structures of
membrane-bound desaturases eventually become available.
Our data also have implications for protein engineering
experiments aimed at converting desaturases into highly regio-
selective and enantioselective hydroxylators. Broun et al.5b have
shown recently that a bifunctional ∆12 desaturase/12-hydroxyl-
ase can be tuned in the direction of dehydrogenation or
hydroxylation by the exchange of only 6 protein residues. We
had previously shown6d using our KIE methodology that a
related plant ∆12 desaturase initiates oxidation at C-12 which
implies that the protein switch controlling chemoselectivity
probably operates after a common initial first step (Scheme 2).
On the basis of the results reported in this paper, one would
then predict that our ω Ϫ 3 desaturase could potentially be
engineered to give a chiral 15-hydroxylated product from a C18
substrate.
316, M+ (methyl 7-thiastearate); m/z 152, (CH (CH ) CH᎐
᎐
3
2 6
CHCH᎐CH )+ (methyl 7-thiaoleate); m/z 150, (CH3–CH2–
᎐
2
CH᎐CH–(CH ) –CH᎐CH–CH᎐CH )+ (methyl 7-thiaisolinole-
᎐
᎐
᎐
2
3
2
ate). Flash chromatography using silica gel (230–400 mesh)
was used to purify substrates. Analytical TLC was performed
using Merck glass plates pre-coated with silica G/UV 254.
Visualization of UV-inactive materials was accomplished by
using a combination of I2 vapour followed by a water spray.
Materials
Unless otherwise stated, all reagents and starting materials were
purchased from Aldrich Chemical Co. and used without purifi-
cation. Tetrahydrofuran (THF) and diethyl ether (Et2O) were
freshly distilled from Na and LAH, respectively. All air- and
moisture-sensitive reactions were performed under N2. Organic
extracts were typically shaken with saturated NaCl and dried
over Na2SO4, and solvents were evaporated on a rotary evapor-
ator. Methyl 7-thiastearate 2 was prepared as previously
described.16
Synthesis of deuterated intermediates
[1,1-2H2]Butan-1-ol. Butanoic acid (4.85 g, 55 mmol) was
added under N2 to a slurry of LiAlD4 (2.51 g, 60 mmol) in Et2O
(25 cm3) at 0 ЊC over 30 minutes. After stirring at RT for 4 h, the
reaction was quenched with the dropwise addition of H2O (10
cm3) followed by HCl (3 M, 20 cm3) and extracted with Et2O
(2 × 25 cm3). The combined organic layers were washed with
HCl (3 M, 30 cm3), dried and evaporated to yield the title com-
pound as a colourless liquid (3.14 g, 75%). For the unlabelled
material: δH 0.91 (3H, t, J 6.5, CH3), 1.36 (2H, sextet, J 7.6,
CH3CH2CH2), 1.51 (2H, quintet, J 7.5, CH2CH2OH), 3.63 (2H,
In summary, we have successfully uncovered the cryptic site
of initial oxidation for an ω Ϫ 3 desaturase which introduces
the third double bond of a central plant lipid—α-linolenic acid
1. The success rate of our KIE approach in elucidating the
cryptoregiochemistry of membrane-bound desaturases remains
100%.
22
t, J 6.5, CH2OH); δC 13.75 (4-C), 18.83 (3-C), 34.66 (2-C),
62.23 (1-C). The title compound showed similar spectral char-
acteristics except for δH 1.47 (2H, br t, J 7.4, CH2CD2OH), 3.63
(absent); δC 18.78 (3-C, upfield γ-isotope shift (0.05 ppm)),
34.45 (2-C, upfield β-isotope shift (0.21 ppm)), 62.23 (1-C,
absent).
Experimental
General
All 1H and 13C NMR spectra were obtained at 200 and 50 MHz
respectively on a Varian XL 200 spectrometer using dilute
CDCl3 solutions. Chemical shifts are expressed in parts per
million (ppm) and are referenced to TMS. J-Values are in Hz.
Deuterium isotope effects on 13C NMR shifts were estimated by
[1,1-2H2]-1-Bromobutane. [1,1-2H2]Butan-1-ol (3.08 g, 34
mmol) was added to a solution of HBr (48%, 7.5 g) and conc.
H2SO4 (1.2 cm3) followed by additional conc. H2SO4 (1.0 cm3).
The solution was refluxed under N2 for 3 h. Distillation directly
from the reaction vessel gave the title compound as a colourless
liquid (3.35 g, 71%). For the unlabelled material: δH 0.94 (3H, t,
J 7.5, CH3), 1.50 (2H, sextet, J 7.0, CH3CH2CH2), 1.83 (2H,
** As has been pointed out on a previous occasion,6b the magnitude of
the large primary deuterium KIE such as that observed for C-15–H
cleavage must be regarded as an estimate since the observed value may
incorporate a small (<10%) α-secondary isotope effect.20 In addition,
partial masking of the “intrinsic” KIE by other enzymic steps in the
catalytic cycle such as substrate binding may also be occurring.21 None
of these considerations affect the conclusions reached in this paper.
22,23
quintet, J 7.3, CH2CH2Br), 3.40 (2H, t, J 6.8, CH2Br); δC
13.17 (4-C), 21.30 (3-C), 33.66 (1-C), 34.76 (2-C). The title
compound showed similar spectral characteristics except for
J. Chem. Soc., Perkin Trans. 1, 2001, 1116–1121
1119

View Article Online
δH 1.81 (2H, br t, J 7.8, CH2CD2Br), 3.40 (absent); δC 21.25
(3-C, upfield γ-isotope shift (0.05 ppm)), 33.66 (1-C, absent),
34.52 (2-C, upfield β-isotope shift (0.24 ppm)).
coupled product, hexadeca-1,15-diene, as judged by GC-MS.
This intermediate was used without further purification.
Crude [9,9-2H2]dodec-1-ene (3.53 g, 21 mmol) was oxidized
with acidic KMnO4, as previously described,18 to give the chain-
shortened carboxylic acid [8,8-2H2]undecanoic acid (3.57 g,
90% est.) as a white solid. This material was reduced with LAH
(see above) to give the corresponding alcohol, [8,8-2H2]-
undecan-1-ol (2.24 g, 68% est. yield) and subsequently treated
with tosyl chloride to give a low melting white solid, [8,8-2H2]-1-
(tolyl-p-sulfonyloxy)undecane (2.91 g, 70% est. yield). For the
known28 unlabelled compound: δH 0.87 (3H, t, J 6.0, CH3), 1.20
(16H, br s, CH3(CH2)8CH2), 1.80 (2H, quintet, J 6.6, CH2-
CH2OTs), 2.43 (3H, s, Ar-CH3), 4.00 (2H, t, J 6.6, CH2OTs),
7.32 (2H, d, ArH), 7.77 (2H, d, ArH); δC 14.06 (11-C), 21.55
(Ar-CH3), 22.62 (10-C), 25.25 (2-C), 28.73, 28.85, 29.25 (8-C),
29.33, 29.43, 29.50, 31.83 (9-C), 70.66 (1-C), 127.79, 129.75,
133.10, 144.58. The labelled tosylate showed similar spectral
characteristics except for δC 22.57 (10-C, upfield γ-isotope shift
(0.05 ppm)), 29.25 (8-C, absent), 31.62 (9-C, upfield β-isotope
shift (0.21 ppm)).
[1,1-2H2]Dec-9-en-1-ol. Dec-9-enoic acid (prepared by Jones
oxidation24 of the commercially available dec-9-en-1-ol) (11.78
g, 69 mmol) was reduced with LiAlD4, as described above, to
give [1,1-2H2]dec-9-en-1-ol as a colourless liquid (9.62 g, 88%).
For the known25 unlabelled material: δH 1.28 (10H, br s,
CH2(CH2)5CH2), 1.54 (2H, quintet, J 6.8, CH2CH2OH), 2.02
(2H, m, CH ᎐CHCH ), 3.62 (2H, t, J 6.4, CH OH), 4.95 (2H,
᎐
2
2
2
m, CH ᎐CH), 5.80 (1H, m, CH ᎐CH); δ 25.71 (3-C), 28.85,
᎐
᎐
2
2
C
29.03, 29.36, 29.42, 32.66 (2-C), 33.75 (8-C), 62.71 (1-C), 114.08
(10-C), 139.09 (9-C). The title compound showed similar spec-
tral characteristics except for δH 1.53 (2H, br t, CH2CD2OH),
3.62 (absent); δC 25.66 (3-C, upfield γ-isotope shift (0.05 ppm)),
32.45 (2-C, upfield β-isotope shift (0.21 ppm)), 62.71 (1-C,
absent).
[1,1-2H2]-1-Bromodec-9-ene. PBr3 (9.98 g, 36 mmol) was
added to a solution of [1,1-2H2]dec-9-en-1-ol (8.24 g, 52 mmol)
in anhydrous Et2O (35 cm3) at 0 ЊC under N2. After reflux (2.5
h), the product was cooled and quenched with H2O (10 cm3).
The organic layer was separated and the aqueous layer was
extracted with Et2O (2 × 20 cm3). The combined organic
extracts were washed, dried (Na2SO4), evaporated and purified
by chromatography on silica, with hexanes as eluent, to give
[10,10-2H2]-1-bromodec-9-ene as a colourless liquid (7.47 g,
65%). For the known26 unlabelled material: δH 1.30 (10H, br s,
CH2(CH2)5CH2), 1.82 (2H, quintet, J 7.0, CH2CH2Br), 2.02
A 1.6 M n-butyllithium (12.5 cm3) reagent was added, via
syringe, to a solution of 6-mercaptohexanoic acid (1.36 g, 9.2
mmol) in 3 : 1 THF–HMPA (28 cm3) under N2 at 0 ЊC. After
stirring at 0 ЊC for 30 min, a solution of [8,8-2H2]-1-(tolyl-p-
sulfonyloxy)undecane (2.79 g, 8.5 mmol) in THF (3 cm3) was
added via syringe and stirred at RT for 24 h. The reaction was
quenched with the addition of H2O (30 cm3), acidified to a pH
of 2 with 3 M HCl (15 cm3) and extracted with hexanes (4 × 30
cm3). The combined organics were washed, dried (Na2SO4) and
evaporated to give crude [15,15-2H2]-7-thiastearic acid (2.14 g,
82% est.) as a white solid. This compound was methylated
with BF3–MeOH (pre-prepared reagent) and purified by flash
chromatography using 4% EtOAc–hexanes to give methyl
[15,15-2H2]-7-thiastearate, [15,15-2H2]-26d (580 mg, 26%) as a
colourless oil which solidified at 20 ЊC. The physical and spec-
tral data of the title compound were similar to those of known
unlabelled 26d except for δC 22.66 (17-C, upfield γ-isotope shift
(0.05 ppm)), 29.32 (15-C, absent), 29.43 (14-C, upfield β-isotope
shift (0.21 ppm)), 31.77 (16-C, upfield β-isotope shift (0.21
ppm)); m/z 318 (M+), 287 ((M Ϫ OCH3)+), 189 ((CH3(CH2)2-
CD2(CH2)7S)+), 129 ((M Ϫ CH3(CH2)2CD2(CH2)7S)+).
(2H, m, CH ᎐CHCH ), 3.41 (2H, t, J 7.0, CH Br), 4.95 (2H,
᎐
2
2
2
m, CH ᎐CH), 5.80 (1H, m, CH ᎐CH); δ 28.16 (3-C), 28.72,
᎐
᎐
2
2
C
28.86, 29.01, 29.28, 32.82 (2-C), 33.78 (8-C), 34.00 (1-C), 114.18
(10-C), 139.08 (9-C). The title compound showed similar
spectral characteristics except for δH 1.81 (2H, br t, J 7.1, CH2-
CD2Br), 3.41 (absent); δC 28.11 (3-C, upfield γ-isotope shift
(0.05 ppm)), 32.58 (2-C, upfield β-isotope shift (0.24 ppm)),
34.00 (1-C, absent).
Synthesis of substrates
The title compounds were prepared using a sequence of well-
established synthetic procedures.
Methyl [16,16-2H2]-7-thiastearate (7-thiaoctadecanoate)
[16,16-2H2]-2. A solution of [1,1-2H2]-1-bromodec-9-ene (6.98 g,
32 mmol) in THF (5 cm3) was added to 1.0 M ethylmagnesium
bromide (50 cm3) followed by 2.0 cm3 of a freshly prepared
dilithium tetrachlorocuprate solution (LiCl (0.21 g, 0.5 mmol)
and CuCl2 (0.34 g, 0.25 mmol) in THF (25 cm3)). The reaction
was stirred for 3 h at 0 ЊC and then quenched with ice (25 g) and
extracted with Et2O (3 × 25 cm3). The combined organic layers
were washed with NH4OH (2 × 25 cm3), Na2S2O3 (1 × 25 cm3),
H2O (2 × 25 cm3), dried over Na2SO4 and evaporated to give
[10,10-2H2]dodec-1-ene as a colourless liquid (4.25 g, 78% est.).
The labelled material showed similar spectral characteristics to
those described above for [9,9-2H2]dodec-1-ene except for
δC 14.12 (12-C, upfield γ-isotope shift (0.05 ppm)), 22.57 (11-C,
upfield β-isotope shift (0.21 ppm)), 32.01 (10-C, absent). This
material was contaminated with a small amount (ca. 20%) of
the starting compound, [1,1-2H2]-1-bromodec-9-ene, as judged
by GC-MS. This intermediate was used without further
purification.
Methyl [15,15-2H2]-7-thiastearate (7-thiaoctadecanoate)
[15,15-2H2]-2. Oct-1-enylmagnesium bromide was prepared by
addition of 8-bromooct-1-ene (5.92 g, 31 mmol) to a mixture of
magnesium (0.76 g, 31 mmol) and I2 (100 mg) in THF (30 cm3)
under N2. After an exothermic reaction had subsided, the Grig-
nard reagent was added, via syringe, to a solution of [1,1-2H2]-
1-bromobutane (4.24 g, 31 mmol) in THF (25 cm3) at –10 ЊC
followed by 1.7 cm3 of a freshly prepared dilithium tetrachloro-
cuprate solution (LiCl (0.21 g, 0.5 mmol) and CuCl2 (0.34 g,
0.25 mmol) in THF (25 cm3)). The reaction was stirred for 3 h at
0 ЊC and then quenched with ice (25 g) and extracted with Et2O
(3 × 25 cm3). The combined organic layers were washed with
NH4OH (2 × 25 cm3), Na2S2O3 (1 × 25 cm3), H2O (2 × 25 cm3),
dried over Na2SO4 and evaporated to give [9,9-2H2]dodec-1-ene
as a colourless liquid (4.01 g, 77% est.). For the unlabelled com-
pound: δH 0.87 (3H, t, J 6.7, CH3), 1.25 (16H, br s, CH3(CH2)8-
CH ), 2.02 (2H, m, CH CH᎐CH ), 4.95 (2H, m, CH᎐CH ), 5.80
Crude [10,10-2H2]dodec-1-ene (4.08 g, 24 mmol) was oxidized
with acidic KMnO4, as previously described,18 to give the
chain-shortened carboxylic acid [9,9-2H2]undecanoic acid
(4.25 g, 94% est.) as a white solid. This material was reduced
with LiAlH4 (see above) to give the corresponding alcohol
[9,9-2H2]undecan-1-ol (2.43 g, 61% est.) and subsequently
treated with tosyl chloride to give a low melting white solid,
[9,9-2H2]-1-(tolyl-p-sulfonyloxy)undecane (2.49 g, 57% est.).
The labelled material showed similar spectral characteristics to
᎐
᎐
2
2
2
2
27
(1H, m, CH᎐CH ); δ
14.17 (12-C), 22.78 (11-C), 29.06,
᎐
2
C
29.27, 29.46, 29.63, 29.73, 29.74 (8-C), 32.03 (10-C), 33.93
(3-C), 114.12 (1-C), 139.25 (2-C); m/z 168 (M+). The labelled
material showed similar spectral characteristics except for
δC 22.74 (11-C, upfield γ-isotope shift (0.04 ppm)), 29.53
(8-C, upfield β-isotope shift (0.21 ppm)), 31.82 (10-C, upfield
β-isotope shift (0.21 ppm)); m/z 170 (M+). This material was
contaminated with a small amount (ca. 10%) of the homo-
1120
J. Chem. Soc., Perkin Trans. 1, 2001, 1116–1121

View Article Online
those described above for [8,8-2H2]-1-(tolyl-p-sulfonyloxy)-
undecane except for δC 14.01 (11-C, upfield γ-isotope shift (0.05
ppm)), 22.41 (10-C, upfield β-isotope shift (0.21 ppm)), 29.04
(8-C, upfield β-isotope shift (0.21 ppm)), 31.83 (9-C, absent).
[9,9-2H2]-1-(Tolyl-p-sulfonyloxy)undecane (2.32 g, 7 mmol)
was reacted with 6-mercaptohexanoic acid (1.18 g, 8 mmol), as
described above, to give crude [16,16-2H2]-7-thiastearic acid
(1.78 g, 85% est.) as a white solid. This compound was methyl-
ated with BF3–MeOH (pre-prepared reagent) and purified by
flash chromatography using 4% EtOAc–hexanes to give methyl
[16,16-2H2]-7-thiastearate, [16,16-2H2]-2 (466 mg, 1.5 mmol,
25%) as a colourless oil which solidified at 20 ЊC. The physical
and spectral data of the title compound were similar to those
of known unlabelled 26d except for δC 14.07 (18-C, upfield
γ-isotope shift (0.05 ppm)), 22.50 (17-C, upfield β-isotope shift
(0.21 ppm)), 29.16 (15-C, upfield β-isotope shift (0.21 ppm)),
31.98 (16-C, absent); m/z 318 (M+), 287 ((M Ϫ OCH3)+), 189
((CH3CH2CD2(CH2)8S)+), 129 ((M Ϫ CH3CH2CD2(CH2)8S)+).
References
1 K. Faber, Biotransformations in Organic Chemistry, Springer, Berlin,
1997, p. 208.
2 H. Joo, Z. L. Lin and F. H. Arnold, Nature, 1999, 399, 670.
3 B. J. Wallar and J. D. Lipscomb, Chem. Rev., 1996, 96, 2625.
4 J. Shanklin and E. B. Cahoon, Annu. Rev. Plant Physiol. Plant Mol.
Biol., 1998, 49, 611.
5 (a) E. B. Cahoon, Y. Lindqvist, G. Schneider and J. Shanklin, Proc.
Natl. Acad. Sci. USA, 1997, 94, 4872; (b) P. Broun, J. Shanklin,
E. Whittle and C. Somerville, Science, 1998, 282, 1315.
6 (a) P. H. Buist and D. M. Marecak, J. Am. Chem. Soc., 1992, 114,
5073; (b) P. H. Buist and B. Behrouzian, J. Am. Chem. Soc., 1996,
118, 6295; (c) P. H. Buist, B. Behrouzian, K. A. Alexopoulos,
B. Dawson and B. Black, J. Chem. Soc., Perkin Trans. 1, 1997, 2617;
(d) P. H. Buist and B. Behrouzian, J. Am. Chem. Soc., 1998, 120, 871;
and references cited therein.
7 J. A. Broadwater, B. G. Fox and J. A. Haas, Fett/Lipid, 1998, 100,
103.
8 I. Navarro, I. Font, G. Fabrias and F. Camps, J. Am. Chem. Soc.,
1997, 119, 11335 and references cited therein.
9 (a) R. H. Austin, H.-K. Chang, G. J. Zylstra and J. T. Groves, J. Am.
Chem. Soc., 2000, 122, 11747; (b) H. Fu, M. Newcomb and C. H.
Wong, J. Am. Chem. Soc., 1991, 113, 5878.
Incubation experiments using S. cerevisiae
10 G. K. Cook and J. M. Mayer, J. Am. Chem. Soc., 1994, 116, 1855 and
references cited therein.
To determine the cryptoregiochemistry of the FAD3 desaturase,
the pRS131/MKP-o strain of S. cerevisiae containing a plasmid
with the B. napus fad3 gene under the control of the yeast
GAL 1 promoter15 was used. Cultures were grown in minimal
medium lacking uracil and containing 2% galactose (CM
gal-ura) supplemented with 1 : 1 mixtures of 2 : [15,15-2H2]-2 or
2 : [16,16-2H2]-2 (100 mg LϪ1) and Tergitol (Type NP-40, 0.1%
[v/v]) at 20 ЊC for 3 days followed by 15 ЊC for 3 days. The
plasmid vector pSE936 in the yeast strain MKP-o was used as a
control.
Fatty acid methyl esters (FAME) were prepared from yeast
cultures using methanolic HCl as described previously.15 To
facilitate GC-MS analysis in the isotope effect experiments, the
FAME were fractionated on a Hewlett Packard 1100 Series
HPLC system (Hewlett Packard, Palo Alto, USA) equipped
with an auto-sampler, solvent degasser, quaternary pump,
column heater and diode array detector (monitoring at 205 and
254 nm), which were all controlled by HP Chemstation soft-
ware. The samples were separated on a series of two 4.6 × 125
mm Whatman Partisphere C-18 reverse phase columns
(Whatman Inc., Clifton, NJ), eluting at a rate of 1 mL minϪ1
with 5 mL acetonitrile followed by a 0% to 30% acetone gradi-
ent (15 mL). Fractions eluting from 2 to 7 min (0.5 mL) were
concentrated and analyzed by GC to confirm fractions contain-
ing thia fatty acids. These were pooled for MS analysis.
11 (a) G. J. Schroepfer and K. Bloch, J. Biol. Chem., 1965, 240, 54; (b)
A. Pinilla, F. Camps and G. Fabrias, Biochemistry, 1999, 38, 15272;
(c) J. L. Abad, F. Camps and G. Fabrias, Angew. Chem., Int. Ed.,
2000, 39, 3279; (d) J. E. Baenziger, I. C. P. Smith and R. J. Hill,
Chem. Phys. Lipids, 1990, 54, 17; (e) L. Fauconnot and P. H. Buist,
J. Org. Chem., 2001, 66, 1210; ( f ) D. Meesapyodsuk, D. W. Reed,
C. K. Savile, P. H. Buist and P. S. Covello, Biochemistry, 2000, 39,
11948; (g) A. Rahier, Biochemistry, 2001, 40, 256; (h) A. E. Rettie,
M. Boberg, A. W. Rettenmeier and T. A. Baillie, J. Biol. Chem.,
1988, 263, 13733; (i) H. G. Enoch, A. Catala and P. Strittmatter,
J. Biol. Chem., 1976, 251, 5095.
12 M. McConn and J. Browse, Plant Cell, 1996, 8, 403.
13 P. J. Nestel, Annu. Rev. Nutr., 1990, 10, 149.
14 M. M. Peyou-Ndi, J. L. Watts and J. Browse, Arch. Biochem.
Biophys., 2000, 376, 399.
15 D. W. Reed, U. A. Schäfer and P. S. Covello, Plant Physiol., 2000,
122, 715.
16 P. H. Buist, H. G. Dallmann, R. R. Rymerson, P. M. Seigel and
P. Skala, Tetrahedron Lett., 1987, 28, 857.
17 M. Tamura and J. Kochi, Synth. Commun., 1971, 303.
18 A. P. Krapcho, J. R. Larson and J. M. Eldridge, J. Org. Chem., 1977,
42, 3749.
19 L. Melander and W. H. Saunders, in Reaction Rates of Isotopic
Molecules, R. E. Krieger Publishing, Florida, 1987, pp. 91–128.
20 J. P. Jones and W. F. Trager, J. Am. Chem. Soc., 1987, 109, 2171.
21 D. Northrup, Biochemistry, 1975, 14, 2644.
22 A. Ejchart, Org. Magn. Reson., 1980, 13, 368.
23 T. D. Brown, PhD thesis, University of Utah, 1965.
24 K. Bowden, I. M. Heilbron, E. R. H. Jones and B. C. L. Weedon,
J. Chem. Soc., 1946, 39.
The position of newly introduced double bonds in the
desaturated products was determined by preparing the GC-MS
analysis of fatty acyl diethylamides as described previously.29a,b
25 C. S. Subramanian, P. J. Thomas, V. R. Mamdapur and M. S.
Chadha, Tetrahedron Lett., 1978, 495.
26 D. K. Johnson, J. Donohoe and J. Kang, Synth. Commun., 1994, 24,
1557.
Acknowledgements
27 (a) S. Rang, E. Lippmaa, T. Pehk and O. Eisen, Eesti NSV Tead.
Akad. Toim. Keem. Geol., 1968, 17, 294; (b) E. Lippmaa, S. Rang,
O. Eisen and T. Pehk, Eesti NSV Tead. Akad. Toim. Keem. Geol.,
1967, 16, 351.
28 S. Krishnamurthy, J. Org. Chem., 1980, 45, 2551.
29 (a) R. Nilsson and C. Liljenberg, Phytochem. Anal., 1991, 2, 253;
(b) D. J. Carrier, J. E. Cunningham, L. R. Hogge, D. C. Taylor and
D. I. Dunstan, J. Chromatogr., A., 1995, 715, 317.
This work was supported by the National Research Council of
Canada International Women in Engineering and Science
Program in association with the Government of Thailand
(D. M.) and the Natural Sciences and Engineering Research
Council of Canada (P. H. B.). The GC-MS data were obtained
with the assistance of Mr Steve Ambrose.
J. Chem. Soc., Perkin Trans. 1, 2001, 1116–1121
1121