Enzymatic kinetic resolution of hydroxystearic acids: A combined experimental and molecular modelling investigation

Article abstract of DOI:10.1016/j.molcatb.2012.06.017

Enantioenriched 7-, 8-, 9-, and 10-hydroxystearic acids (HSA) were obtained, for the first time, by kinetic resolution of their racemates with lipases CALB and PS, in the presence of vinyl acetate. Among them, the best results were obtained for 7-HSA and 9-HSA, whose enantiomeric excess was around 55%. The same resolutions carried out on the hydroxy esters completely failed. For the acid substrates neither the Kazlauskas' rule nor the 3D-QSAR model could be applied, since both models are focused on the CALB alcohol-pocket evaluation and not on the acyl-pocket one. Therefore, a semiquantitative approach was used, whose results were in accordance with our findings, as far as the absolute configuration of the product is concerned.

Full text of DOI:10.1016/j.molcatb.2012.06.017

Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Enzymatic kinetic resolution of hydroxystearic acids: A combined experimental
and molecular modelling investigation
Cynthia Eberta, Fulvia Fellugaa, Cristina Forzatoa, Marco Foscatoa, Lucia Gardossia,
Patrizia Nittia,∗, Giuliana Pitaccoa, Carla Bogab,∗∗, Paolo Caruanab, Gabriele Michelettib,
Natalia Calonghic, Lanfranco Masottic
a Dipartimento di Scienze Chimiche e Farmaceutiche, Università di Trieste, via Licio Giorgieri 1, I-34127 Trieste, Italy
b Dipartimento di Chimica Organica ‘A. Mangini’ Alma Mater Studiorum, Università di Bologna Viale Risorgimento, 4-40136 Bologna, Italy
c Dipartimento di Biochimica ‘G. Moruzzi’ Alma Mater Studiorum, Università di Bologna Via Irnerio, 48-40126 Bologna, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 17 May 2012
Received in revised form 28 June 2012
Accepted 30 June 2012
Available online 10 July 2012
Enantioenriched 7-, 8-, 9-, and 10-hydroxystearic acids (HSA) were obtained, for the first time, by kinetic
resolution of their racemates with lipases CALB and PS, in the presence of vinyl acetate. Among them,
the best results were obtained for 7-HSA and 9-HSA, whose enantiomeric excess was around 55%. The
same resolutions carried out on the hydroxy esters completely failed. For the acid substrates neither the
Kazlauskas’ rule nor the 3D-QSAR model could be applied, since both models are focused on the CALB
alcohol-pocket evaluation and not on the acyl-pocket one. Therefore, a semiquantitative approach was
used, whose results were in accordance with our findings, as far as the absolute configuration of the
product is concerned.
Keywords:
Hydroxystearic acids
Enzymatic resolution
CALB
© 2012 Elsevier B.V. All rights reserved.
Molecular recognition
Enantioselectivity
1. Introduction
oleic acid can be converted into 10-HSA with a variety of bacte-
rial species or yeasts [16], (S)-dimorphecolic acid present in large
Hydroxy fatty acids and their derivatives belong to a class of
compounds of growing interest in many areas. Apart from being
essential molecules in many biological processes [1,2], they are also
used in a variety of industrial processes [3–10] and more recently
also in material chemistry [11].
In connection with their biological activities, the most studied
members of the series have been 9- and 10-hydroxystearic acids
(9- and 10-HSA), for their natural negative regulatory activity of
tumour cell proliferation [12]. Recently [13], 9-HSA has been found
to play an important role in cancerogenesis of HT29, a human
colon adenocarcinoma cell line, acting as a competitive inhibitor of
HDAC1 (histone deacetylase 1) isoform [14]. Furthermore, studies
of molecular docking [14] have shown a more favourable bind-
ing energy for the HDAC1-(R)-9-HSA complex, as compared to the
(S) complex. Therefore the availability of hydroxystearic acids in
enantiopure forms seems an attractive challenge.
amount in the seeds of genus Dimorphotheca, can be a precursor
of (R)-9-HSA [15] while d-ricinoleic acid, obtained from castor oil,
can be a precursor of (R)-12-HSA [17]. Less available from natural
sources are precursors of 8-HSA, namely isanolic acid [15,18] and
laetisaric acid [19], whereas no useful natural precursor of 7-HSA
is known. Obviously, since the choice of the plant must take into
account both its availability and the relative amount of the desired
compound in the seed oil, an enantioselective synthesis might be
envisaged as a preferable route.
Enantioselective chemical syntheses of hydroxystearic acids or
their unsaturated precursors have been reported only in a few cases
[20] and require multi-step synthetic strategies. On the other hand,
it is well-known how difficult it is to achieve a high degree of enan-
tioselectivity in the reduction of carbonyl compounds with a quasi
symmetric long-chain substitution [21]. The same difficulties were
encountered for biocatalysed reductions that gave good results only
for short-chain ␥- and ␦-ketoacids [22,23].
In general, their natural precursors are unsaturated hydroxy
fatty acids contained in seeds of higher plants [15]. For example,
All these results prompted us to investigate the possibility of
preparing a few members of the series in optically pure form, by
the use of enzymes. The scope was also to explore how the stere-
ochemical course could depend on the position of the hydroxyl
group along the chain, especially when situated far from the termi-
nus of the chain. In this preliminary study, Lipase B from Candida
∗
Corresponding author. Tel.: +39 0 405583923; fax: +39 0 405583903.
Corresponding author. Tel.: +39 0 512093616.
E-mail addresses: pnitti@units.it (P. Nitti), carla.boga@unibo.it (C. Boga).
∗∗
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2012.06.017

C. Ebert et al. / Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
39
MgBr
SOCl2
Et2O,
HO
O
Cl
O
m-1
O
n
n
m
n
THF, -78°C, 5 min
O
OCH3
O
OCH3
O
OCH3
1c
2b,c
3a–d
NaBH4
CH3OH
r.t.
a: n = 5, m = 10
b: n = 6, m = 9
c: n = 7, m = 8
d: n = 8, m = 7
KOH/CH3OH
n COOH
m
n COOCH3
m
OH
OH
5a–d
4a–d
Scheme 1. Synthesis of isomeric HSAs 5a–d.
antarctica (CALB) and lipase from Pseudomonas cepacia were cho-
sen since the experimental outcomes could be compared with
theoretical studies. In fact, the resolution of racemic compounds
by CALB is one of the most studied biocatalytic systems [24]. In
hydrolysis/synthesis of esters, the enantiopreference is predicted
by Kazlauskas’ rule [25] that assesses that in the case of a secondary
alcohol with two substituents of different steric encumbrance
the (R)-enantiomer reacts preferentially. This was confirmed by
molecular modelling studies [24–28] and, more recently, by three-
dimensional structure-activity models (3D-QSAR) [29a,29b]. Being
aware of the challenge represented by the resolution of these
particular bulky and quasi symmetric secondary alcohols, a prelim-
inary computational analysis was performed. Docking, molecular
dynamic simulations and 3D-QSAR models were used to verify
the ability of the enzyme to recognize and transform these sub-
strates. A comparison was also made with the lipase from P. cepacia.
The results emerged from the computational analysis were also
supported by the experimental data, which confirmed how the sec-
ondary hydroxyl group on the fatty acid chain is in an unfavourable
position both in terms of reactivity and in terms of enzyme enan-
tiodiscrimination.
to evaluate a possible energy difference in the tetrahedral reaction
intermediates, by comparing, for each enantiomer, the distribution
of the molecular dynamic (MD) poses stabilized by a certain num-
ber of H-bonds able to stabilize the oxy-anion hole and to allow
the evolution to the subsequent reaction steps. In fact, it is known
that the R- and S-enantiomers have different orientations to fulfil
the necessary network of hydrogen bonds in transition state when
reacting with CALB [32]. In the present study the computational
analysis was performed in four steps (see also Section 4): (1) build-
ing up the tetrahedral intermediates inside the CALB active site; (2)
conformational search to identify the lower energy conformation of
substrate acyl chain; (3) equilibration by molecular dynamic simu-
lation; (4) scoring of tetrahedral intermediates by analysing H-bond
pattern. An analogous approach was then followed also for lipase
from P. cepacia [33]. When one of the two enantiomers is stabi-
lized at a higher extent, a difference in activation energy between
the reactions of the two reactions is expected, which ultimately
translates into enantiodiscrimination. As an example, Part (C) of
Fig. 1 shows the H-bond profiles for the tetrahedral intermediate
(i.e. transition state analog) of enantiomers of 1,2-dimethylpropyl
octanoate, whose resolution with CALB leads to E > 700 [28]. It
is evident that the (R)-enantiomer presents a higher number of
“poses” stabilized by 3–5 H-bond, whereas the slow reacting (S)-
enantiomer has a higher chance to be stabilized by only one or
two H-bonds. Therefore, statistically speaking, there is a higher
number of conformational poses of the tetrahedral intermediate
of (R)-enantiomer having lower energy and being more reactive.
Data reported in Fig. 1A and B show that the poses of tetrahe-
dral intermediates of both enantiomers of ester 4d establish similar
numbers of H-bonds inside the active sites of the two lipases here
considered. Therefore, the tetrahedral intermediates of the enan-
tiomers will be stabilized at a similar extent, leading to comparable
reaction rates. As a conclusion, enantiodiscrimination seems unfea-
sible for substrate 4d by using either CALB or lipase from P. cepacia.
Therefore, this preliminary study clearly indicates that the “nat-
ural” reactivity of the carboxyl group cannot be exploited for the
enantioresolution of the HSA ester 4d.
2. Results and discussion
Racemic 7-, 8-, 9-, and 10-hydroxystearic acids (HSA) 5a–d were
synthesized by chemical reduction of the corresponding 7-, 8-, 9-,
and 10-ketoesters 3a–d (Scheme 1), followed by hydrolysis under
basic conditions [12c,30]. Compounds 3a and 3d are commercially
available while 3b and 3c were prepared from 2b (commercially
available) and 2c, respectively, by addition of the appropriate Grig-
nard reagent. Compound 2c was in turn prepared from azelaic acid
monomethyl ester 1c and thionyl chloride [31].
Since two functionalities are present in these molecules, two dif-
ferent biocatalytic approaches were considered, the first involving
hydrolysis of the ester group in the hydroxyesters and the second
involving acylation of the hydroxyl group in the hydroxystearic
acids, in organic solvent.
In particular, the resolutions of the hydroxyester 4d and the
hydroxystearic acids 5a–d were first studied by computational
analysis to verify the feasibility of their enantiodifferentiation by
CALB and by lipase from P. cepacia.
2.2. Computational analysis: CALB substrate recognition and
reactivity on the secondary OH group in 7-, 8-, 9-, and 10- HSA
5a–d
2.1. Computational analysis: CALB enantiodiscrimination by
exploiting the reactivity of the acyl group in 10-HSA methyl ester
4d
We recently constructed a 3D-QSAR model able to describe and
predict the enantioselectivity of CALB [29]. The application of the
3D-QSAR model previously developed implies a number of com-
putational steps: (a) calculation of the active conformers of the
tetrahedral intermediates for all enantiomers of 5a–d by molecular
Trying to describe how the acylic pocket of CALB recognizes the
enantiomers of the ester 4d, a semi-quantitative approach was used

40
C. Ebert et al. / Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
Fig. 1. Distribution of poses of the tetrahedral intermediates of enantiomers stabilized by a defined number of H-bonds. (A) Tetrahedral intermediate of 4d in CALB; (B)
tetrahedral intermediate of 4d in lipase from Pseudomonas cepacia; (C) tetrahedral intermediate of 1,2-dimethylpropyl octanoate in CALB. X axis represents the number of
H-bonds established by the substrate inside the acylic pocket. Y axis expresses the frequency of a specific molecular dynamic pose characterized by a certain number of
stabilizing H-bonds.
dynamics simulations; (b) generation of the molecular descrip-
tors for the pairs of enantiomers by means of GRID analysis and
DMIF calculation; (c) calculation of the E value from the 3D-QSAR
regression model (PLS partial least squares – projection to latent
structures) [29].
residues Glu188 and Ile189, on the contrary, would lead to a wrong
displacement of Asp187 from the catalytically productive confor-
mation (lower simulations of Fig. 2). Therefore, these discrepant
observations cannot lead to any conclusive prevision of CALB reac-
tivity to HSAs and indicate that Kazlauska’s rule cannot be applied
Initially, it was necessary to model the tetrahedral intermedi-
ates for enantiomers of substrates 5a–d. However, a preliminary
study of docking of the substrate ground state into the alcohol
pocket, by using the docking algorithm AlphaPMI, was unable to
produce reliable poses consistent with catalysis. HSAs are charac-
terized by a high number of rotatable bonds that result into an
exceeding conformational freedom. Because of the various rotat-
able bonds, the substrates got coiled in the funnel shaped active
site, thus producing no pose. Attempts to modify the placement
algorithm failed basically because, in agreement with the CALB
biological role, the carboxylic group of 5a–d is better accommo-
dated in the acylic pocket rather than in the alcohol pocket (see
Fig. S1 in Supplementary Material).
To overcome these limitations and to model reliable tetrahedral
intermediates, a pharmacophoral based approach (see Section 4)
was applied by placing C1-decarboxylated derivatives of HSA acyl
esters into the active site and by imposing a conformation compat-
ible with catalysis (Fig. S2 in Supplementary Material). Afterwards
the structure of the models was completed by reintroducing the
carboxylic moiety on C1 and constructing the tetrahedral interme-
diates from the acetic ester moiety.
A molecular dynamic simulation was performed in order to relax
the tetrahedral intermediates and evaluate the structures to be
used in the 3D-QSAR model (300 ps, NVT, freezing atoms more than
˚
10 A far away from the TI).
It must be noted that both chains bounded to the alcohol moiety
are too long and bulky to be accommodated freely inside the stere-
ospecificity pocket. Interestingly, the MD relaxation showed some
enlargement and conformational adaptations of the active site fun-
nel for accommodating such bulky substrates. Although the short
MD simulation did not allow any conclusive analysis of the con-
formational behaviour of CALB in the presence of HSA substrates,
the observation suggests that a movement of ␣-10 helix [34], might
enable a better substrate binding (upper simulation of Fig. 2). How-
ever, the consequent displacement of the loop corresponding to
Fig. 2. Displacement of Asp187 and partial unfolding of ␣-10 helix enlarges the
alcohol pocket of CALB active site. Crystallographic structure of CALB (1LBT) is in
blue and final MD snapshot in red. The small frame on the right (blue) shows the
productive position of the three aminoacids of the catalytic triad, with Asp187 form-
ing a H-bond with His224. The frame on the left (red) illustrates the displacement
of Asp187 as a result of the fitting of the bulky hydroxyacyl moiety inside the stere-
ospecificity pocket. Conversely, Asp187 cannot interact with His224 in a productive
way. (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of the article.)

C. Ebert et al. / Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
41
Lipase
OAc
S
R
m
+
m
n
n
COOH
COOH
m
n
COOH
OH
5a-d
OH
OAc
6a-d
unreacted 5a-d
CH2N2
a: n = 5, m = 10
b: n = 6, m = 9
c: n = 7, m = 8
d: n = 8, m = 7
m
n
+
COOCH3
m
n
COOCH3
OH
OAc
7a-d
4a-d
1. H3O+
OAc
2. CH2N2
OH
O
DCC
m
n
COOCH3
O
O
AcO
Fig. 3. Predictivity of the model in terms of experimental versus predicted E values.
Objects taken from the training set of the PLS model are represented by squares. The
performance of the model towards four HSAs, 5a–b, is highlighted by crosses.
8a-d
Scheme 2. Lipase catalysed acylation of hydroxystearic acids 5a–d.
not be made for 5a and 5b and their absolute configurations were
tentatively assigned as (S) for analogy.
to HSA acyl esters. By applying the cited 3D-QSAR approach to
5a–d, the model predicted an enantiopreference towards the (R)
enantiomer. However, the indications obtained from the 3D-QSAR
model induced to test this type of reactivity looking for, at least,
some enantiomeric enrichment.
Experimental data clearly show that CALB is able to acylate the
OH group of HSAs despite a very low catalytic efficiency, as pre-
dicted by docking studies. Since the substrates are too bulky for
being accommodated in the stereospecific alcohol pocket, some
induced fit of CALB might be involved in the enzyme-substrate
recognition, although extensive MD simulation would be necessary
to explore this hypothesis in further detail.
2.3. Experimental results
As for enantioselectivity, modest enantiodiscrimination was
observed. The indications gained from the 3D-QSAR model were
correct in term of prediction of the fast reacting enantiomer but a
clear overestimation of enantioselectivity was done by the model
(Fig. 3) due to the diversity between acetoxystearic acids and the
data set used to train the 3D-QSAR model.
Lipase catalysed hydrolysis of methyl 10-hydroxystearate 4d.
As a confirmation of the theoretical evaluations, both lipases
are able to catalyse the hydrolysis of methyl 10-hydroxystearate
4d (Table 1) with acceptable reaction rate but with low enantiose-
lectivity, as shown by their very low E values.
As it can be seen in Table 2, calculated and measured conversion
values were similar with the exception of those reported for 5d in
its reaction with CALB. As expected on the basis of the substrate
structures, E values were low for all reactions and consequently
the optical purity of the products were not high, ranging from 16
to 56%.
Acylation reactions carried out with lipase PS were less satis-
factory and only two results are reported in Table 2, namely those
relative to compounds 5c and 5d.
Interestingly, the acylation reaction carried out on the esters
4a–d did not proceed, most probably because the enzyme recog-
nizes the fatty ester leading to an acyl enzyme that, however, has
no nucleophile available for being transformed at a reasonable rate.
Indeed, the water activity is too low. As a matter of fact, the ester is
a more active acylating agent as compared to the carboxylic acid.
When the HSA enters the acylic pocket it hardly react with the
catalytic Ser105, so that it has a chance to leave the acylic pocket
unreacted and then to interact with the enzyme according different
modalities.
It is likely that also the ester fits preferentially inside the acylic
pocket (as the acid does). In that case the catalytic Ser105 gets acy-
lated and ethanol is released but no further nucleophile is available
in the reaction environment to attack the acyl-enzyme. In fact, the
HAS is too bulky to be accommodated inside the active site next the
acyl-enzyme. Therefore the acyl-enzyme is not transformed. As a
consequence, the HSA ester is not available for being acylated on
the OH group.
2.4. Lipase catalysed acylation of hydroxystearic acids 5a–d with
CALB and PS
The experimental results of lipase catalysed acylations carried
out on the HSAs 5a–d using lipase B from C. antarctica (CALB) and
lipase from P. cepacia “Lipo P. Cepacia” (PS) with vinyl acetate as
the acylating agent are reported in Scheme 2 and Table 2.
Reactions were monitored by 1H NMR, following the appear-
ance of a quintet at about 4.8 ppm relative to the CHOAc protons
of the acetoxy acids 6a–d. The e.e.’s of both 5a–d and 6a–d were
determined following a complex procedure shown in Scheme 2.
The reaction mixtures containing the unreacted hydroxyacids 5a–d
and the reaction products 6a–d were treated with diazomethane
to esterify both carboxylic groups. The resulting derivatives 4a–d
and 7a–d, respectively were separated by preparative TLC and
separately derivatized with (R)-(−)-O-acetylmandelic acid, the lat-
ter compound after hydrolysis and esterification (Scheme 2). The
resulting diastereomeric acetylmandelate derivatives 8a–d were
then analysed by 600 MHz 1H NMR to determine the enantiomeric
excesses of the parent compounds 5a–d and 6a–d. The absolute
configurations of 8a and 8d, derived from the unreacted hydroxy-
acids 5c and 5d, were (9S,2ꢀR) and (10S,2ꢀR), respectively [16a,36].
In fact, these latter compounds were the major components in their
respective diastereomeric mixtures. As a consequence, the parent
hydroxyacids 5c and 5d are (S)-9-HSA and (S)-10-HSA and the ace-
toxyacids 6c and 6d must be (R). An analogous determination could

42
C. Ebert et al. / Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
Table 1
Lipase catalysed hydrolyses of methyl 10-hydroxystearate 4d.
Enzyme
Time (d)
Conv. (%)a
Unreacted substrate 4d
Ec
e.e.s (%)b
CALBd
PS
2.5
2.5
58
50
12
7
1.3
1.2
a
Determined by 1H NMR analysis.
b
Determined by integration of the methoxy signal in the 600 MHz 1H NMR spectrum of diastereomeric (R)-(−)-O-acetylmandelate esters (see infra).
ln[(1−c)(1−e.e. )]
s
c
E =
; Ref. [35].
ln[(1−c)(1+e.e. )]
s
d
Novozym 435 and LipoCALB gave the same results.
3. Conclusions
CALB), ≥2000 U/g (Sprin Technologies S.p.A., Trieste, Italy) and
lipase from P. cepacia “Lipo P. Cepacia” (PS), ≥100 U/g (Sprin Tech-
nologies S.p.A., Trieste, Italy) were used.
The structural anomaly of hydroxystearic acids is the presence
of a chiral carbon on the acyl moiety, which make difficult the appli-
cation of general rules in the prediction of CALB enantiopreference.
As a matter of fact, the acyl moiety is accommodated into the acyl
pocket of CALB which is not responsible for enantiodiscrimination.
Indeed, the Kazlauskas’s rule as well as the 3D-QSAR model pre-
viously developed for predicting CALB enantiodiscrimination are
strictly connected to the structural features of the alcohol pocket
of CALB. Hydroxy fatty acids are acylated very slowly by lipase
from C. antarctica due to the preferred fitting of the carboxyl group
into the acyl pocket of the enzyme. Because of the bulky alcohol
moiety, the stereospecific pocket cannot host the substrates and
thus these secondary alcohols do not follow either Kazlauskas’ rule
or 3D-QSAR models predicting enzyme enantioselectivity. Mod-
elling study illustrated clearly how enantiodiscrimination is not
feasible by exploiting the reactivity of the acylic group, although
reactions are faster through this route. The accommodation of the
acyl chain causes conformational distortion of the alcohol pocket
(Fig. 2). Therefore the substrate recognition occurs according dif-
ferent criteria as compared “normal” alcohols because in the case
of HAS the OH group is hardly positioned in the right pocket.
However, this is the first attempt to get to enantioenriched
hydroxystearic acids by the use of enzymatic resolution. Although
we are conscious of the challenge, these first results seem promis-
ing: two hydroxyacids, namely (R)-7-HSA and (S)-9-HSA were
obtained with about 55% e.e., the former after hydrolysis of its
acetylated derivative. In spite of the poor enantiomeric excesses,
it is our opinion that biological tests would still be worth perform-
ing, in order to have an indication of the influence of the absolute
configuration on the biological activity of the molecules.
Nonanedioic acid, 1-methyl ester (1c), methyl 8-chloro-
8-oxooctanoate (2b), methyl 7-oxooctadecanoate (3a), methyl
10-oxooctadecanoate (3d) and n-decyl magnesium bromide (1.0 M
in diethyl ether) were purchased from Sigma–Aldrich (Milan, Italy).
Methyl 9-oxooctadecanoate (3c) was prepared according to the
literature [12c]. Anhydrous THF was prepared by distillation over
sodium benzophenone ketyl. Thin-layer chromatography (TLC) was
carried out using silica gel precoated on TLC Alu foils from Fluka and
spots were revealed using an aqueous solution of (NH4)6MoO24
(25%), (NH4)4Ce(SO4)4 (1%) in 10% H2SO4 as staining reagent. For
preparative TLC, 20 × 20 silica gel plates (Merck Kieselgel 60F254
)
were used. Flash chromatography was run on silica gel 230–400
mesh ASTM (Kieselgel 60, Merck). Light petroleum refers to the
fraction with b.p. 40–70 ◦C. 1H and 13C NMR spectra were recorded
on a Varian Mercury 400 or an Inova 300 or 600 spectrometer (Var-
ian, Palo Alto, CA). Frequencies are reported in Hz and chemical
shifts in ppm are referred, if not otherwise specified, to the solvent:
CDCl3 (ı = 7.27 ppm for 1H NMR and ı = 77.0 ppm for 13C NMR).
Signal multiplicities were assigned by DEPT experiments. EI and
HRMS mass spectra were recorded using a VG-7070E spectrometer
at an ionization voltage of 70 eV. ESI-MS spectra were recorded on
a WATERS 2Q 4000 instrument. IR spectra were recorded using a
Perkin-Elmer FT-IR MOD.1600 spectrophotometer. Melting points
were measured by a Büchi apparatus and were not corrected.
4.2. Synthesis of substrates
4.2.1. Methyl 8-oxooctadecanoate (3b)
A solution of n-decyl magnesium bromide (1 M in diethyl ether,
15.05 mL) was added, under nitrogen atmosphere, to a solution of
methyl 8-chloro-8-oxooctanoate (2b, 3.11 g, 15.05 mmol) cooled
at −78 ◦C. After 5 min the reaction mixture was treated with water
(20 mL) and the organic layer was separated. The aqueous layer was
extracted with diethyl ether (3× 20 mL). The organic layers were
washed with ‘brine’ then dried over anhydrous MgSO4. After filtra-
tion and removal of the solvent under reduced pressure, the residue
4. Experimental
4.1. General
Lipase from C. antarctica “Novozym 435”, ≥10,000 U/g (Novo
Nordisk A/S, Bagsvaerd, Denmark), lipase from C. antarctica (Lipo
Table 2
Acetylation of hydroxystearic acids 5a–d mediated by CALB and PS.
Enzyme
Substrate
Time (d)
Conv. (%), calc.a
Conv. (%), by 1H NMR
Unreacted substrates
Acetylated acids
Ec
5a–d e.e.s (%)b
6a–d e.e.p (%)b
Novozym 435
Novozym 435
Novozym 435
Novozym 435
PS
5a
5b
5c
5d
5c
5d
14
15
39
56
46
–
13
32
53
37
37
34
10d
24d
54d
24e
30e
16e
56d
38d
42d
28e
–
4
3
4
2.2
4f
1.6
7
5
7
7
7
PS
50
16d
e.e.
a
s
c =
e.e. +e.e.
s
p
Determined by 600 MHz 1H NMR analysis of diastereomeric (R)-(−)-O-acetylmandelate esters of the methyl ester derivatives 8a–d.
b
c
ln[e.e. (1−e.e. )]/(e.e. +e.e.
E =
)) ; Ref. [35].
p
s
p
s
ln[e.e. (1+e.e. )]/(e.e. +e.e.
p
s
p
s
Determined by integration of H-2ꢀ signals of (R)-(−)-O-acetylmandelate derivatives.
d
e
f
Determined by integration of the methoxy group signals of (R)-(−)-O-acetylmandelate derivatives.
ln[(1−c)(1−e.e. )]
s
s
E =
; Ref. [35].
ln[(1−c)(1+e.e. )]

C. Ebert et al. / Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
43
was purified by flash-chromatography (light petroleum/diethyl
4.3. Enzymatic hydrolysis of methyl 10-hydroxystearate 4d
ether 4/1); 1.0 g, (21%) of 3b was obtained. m.p.: 45–46 ◦C (lit.:
[37] 46.4–46.9 ◦C). 1H NMR (400 MHz, CDCl3) ı (ppm): 3.46 (s, 3H,
OCH3), 2.20 (t, J = 7.5 Hz, 2H, CH2CO), 2.19 (t, J = 7.5 Hz, 2H, CH2CO),
2.10 (t, J = 7.5 Hz, 2H, CH2CO), 1.55–1.25 (m, 6 H), 1.25–0.99 (m,
18 H), 0.69 (t, J = 6.8 Hz, 3H, CH3); 13C NMR (100.56 MHz, CDCl3)
ı (ppm): 210.2, 173.4, 50.8, 42.3, 42.1, 33.4, 31.5, 29.2, 29.1, 29.0,
28.90, 28.85, 28.5, 28.4, 24.3, 23.4, 23.1, 22.2, 13.6; MS (m/z), (%):
312 (M+, 2), 281 (15), 239 (9), 199 (8), 184 (56), 171 (76). 154 (61),
144 (70), 129 (100), 111 (67), 97 (56), 83 (70), 69 (61), 55 (78).
To 0.060 g (0.2 mmol) of 4d, 1 mL of acetone, 9 mL of 0.1 M phos-
phate buffer (NaHPO4/KH2PO4) at pH 7.4 and 0.060 g of the lipase
Novozym 435 or Lipo CALB or PS was added. After 2.5 days the aque-
ous suspension was filtered off through a Gooch funnel to recover
the enzyme and the mother liquors were extracted four times with
diethyl ether and dried over anhydrous Na2SO4. Both the hydrox-
yacid 5d and its methyl ester 4d were extracted at pH 7.4 and the
mixture was analysed by 1H NMR. Conversions were determined
by integration of the signals at 3.59 ppm (bs, H-10 of 4d and 5d) and
at 3.68 ppm (s, CH3O of 4d). The ester 4d was separated from the
acid 5d by flash chromatography (petroleum ether/ethyl acetate:
9/1, 25% yield) and the hydroxyl group was esterified with (R)-
(−)-O-acetylmandelic acid [16b]. The enantiomeric excess of the
unreacted substrate was calculated from the 1H NMR (600 MHz)
spectrum of the mixture of the two diastereomers 8d obtained
which agreed with previously reported data [16b,36].
4.2.2. Methyl hydroxyoctadecanoates 4a–d
Compounds 4a–d were prepared from the corresponding ketone
precursor following the procedure reported for the preparation of
4c and 4d [12c]. Characterization data of compounds 4a–d agree
with those reported in the literature [12c,37–39]. The new spectral
data are reported below.
4.2.3. Methyl 7-hydroxyoctadecanoate (4a) [37,38]
1H NMR (600 MHz, CDCl3) ı (ppm): 3.67 (s, 3H, OCH3), 3.62–3.56
(m, 1H, CHOH), 2.32 (t, J = 7.5 Hz, 2H, CH2CO), 1.68–1.60 (m, 2H),
1.51–1.22 (m, 27H), 0.89 (t, J = 7.0 Hz, 3H, CH3); MS (m/z), (%): 296
(0.4), 264 (7), 159 (35), 127 (100), 87 (76).
4.4. Lipase catalysed acylations of hydroxystearic acids 5a–d
To 0.070 g (0.23 mmol) of hydroxystearic acid, 3 mL of vinyl
acetate, 4 mL of diethyl ether and 0.070 g of Novozym 435 or Lipo
P. Cepacia (PS) were added. The mixture was stirred at room tem-
perature for 5–14 days. The suspension was filtered through a
Gooch funnel, the enzyme was washed with diethyl ether and the
solvent was evaporated. Conversions were determined by integra-
tion of the signals at about 3.6 ppm (bs, CHOH of 5a–e) and at
about 4.8 ppm (quintet, CHOAc of 6a–d) of the 1H NMR spectrum
(in CDCl3) of the crude reaction mixture. This latter was dis-
solved in the minimum amount of methanol and esterified ‘in situ’
with diazomethane in diethyl ether using Aldrich® diazomethane-
generator apparatus. After removal of the solvent, preparative TLC
(light petroleum/diethyl ether: 7/3) of the residue allowed to sep-
arate compounds 4 (RF ∼ 0.25) and 7 (RF ∼ 0.7). The related scraped
silica gel bands were suspended and stirred in methanol, giving,
after filtration and removal of the solvent, compounds 4 and 7. The
hydroxy group of the methyl hydroxyester (4a–d) thus obtained
was esterified with (R)-(−)-O-acetylmandelic acid [16b] to deter-
mine the enantiomeric excess of the unreacted substrate by 1H
NMR. The methyl acetoxyoctadecanoates (7a–d) were dissolved
in methanol (1–2 mL) and treated with an excess (1.0–2.0 mL) of
10% KOH/CH3OH. The solution was magnetically stirred at room
temperature and monitored by TLC (light petroleum/diethyl ether:
7/3) until complete disappearance of the starting material. The
flask was immersed in an ice-bath and the solution was acidi-
fied by adding dropwise 37% aq. HCl solution. After filtration, the
methanol was partially removed and the residue containing the
hydroxyacid was treated with diazomethane in diethyl ether. The
solvent was removed and the residue was treated with (R)-(−)-
O-acetylmandelic acid following the reported procedure [16b] to
determine the enantiomeric excess of 6a–d by 1H NMR. The new
spectral data for compounds 7a–d and 8a,b are reported below,
while spectral data for 8c and 8d are consistent with those reported
in the literature [16b,36].
4.2.4. Methyl 8-hydroxyoctadecanoate (4b) [37–39]
M.p.: 54–55 ◦C (lit.: [39a] 54.5–55.5 ◦C). 1H NMR (400 MHz,
CDCl3) ı (ppm): 3.65 (s, 3H, OCH3), 3.60–3.51 (m, 1H, CHOH), 2.29
(t, J = 7.6 Hz, 2H, CH2CO), 1.68–1.16 (m, 29H), 0.87 (t, J = 6.7 Hz, 3H,
CH3); MS (m/z), (%): 296 (0.4), 264 (3), 173 (35), 141 (100).
4.2.5. Hydroxyoctadecanoic acids 5a–d
Compounds 5a–d were prepared from the corresponding
methyl ester following the procedure reported [14] for the
preparation of 9-hydroxylstearic acid 5c. Characterization data
of compounds 5a–d agree with those reported in the literature
[12c,37]. The new spectral data are reported below.
4.2.6. 7-Hydroxyoctadecanoic acid (5a) [40]
1H NMR (400 MHz, CDCl3) ı (ppm): 3.70–3.52 (m, 1H, CHOH),
2.37 (t, J = 7.6 Hz, 2H, CH2CO), 1.74–1.59 (m, 2H), 1.53–1.21 (m,
27H), 0.89 (t, J = 6.7 Hz, 3H, CH3); 13C NMR (100.56 MHz, CDCl3) ı
(ppm): 177.4, 71.9, 37.5, 37.2, 33.5, 31.9, 29.69, 29.66, 29.63 (two
overlapping signals), 29.61, 29.3, 29.1, 25.6, 25.3, 24.6, 22.7, 14.1.
4.2.7. 8-Hydroxyoctadecanoic acid (5b) [37,41]
1H NMR (400 MHz, CDCl3) ı (ppm): 3.63–3.55 (m, 1H, CHOH),
2.35 (t, J = 7.4 Hz, 2H, CH2CO), 1.69–1.59 (m, 2H), 1.51–1.22 (m,
27H), 0.89 (t, J = 6.7 Hz, 3H, CH3); 13C NMR (100.56 MHz, CDCl3) ı
(ppm): 179.3, 72.0, 37.5, 37.3, 33.9, 31.9, 29.7, 29.6 (three overlap-
ping signals), 29.3, 29.25, 29.0, 25.6, 25.4, 24.6, 22.7, 14.1.
4.2.8. 9-Hydroxyoctadecanoic acid (5c) [37]
1H NMR (400 MHz, CDCl3) ı (ppm): 3.66–3.54 (m, 1H, CHOH),
2.37 (t, J = 7.8 Hz, 2H, CH2CO), 1.69–1.60 (m, 2H), 1.60–1.23 (m,
27H), 0.89 (t, J = 7.0 Hz, 3H, CH3); 13C NMR (100.56 MHz, CDCl3)
ı (ppm): 178.7, 72.0, 37.5, 37.4, 33.5, 31.9, 29.7, 29.65, 29.58, 29.5,
29.3, 29.2, 29.0, 25.7, 25.5, 24.7, 22.7, 14.1.
4.4.1. Methyl 7-acetoxyoctadecanoate (7a) [38a,38d,42]
1H NMR (400 MHz, CDCl3) ı (ppm): 4.85 (quintet, 6.7 Hz, 1H),
3.67 (s, 3H, OCH3), 2.30 (t, J = 7.7 Hz, 2H), 2.04 (s, 3H, CH3CO),
1.67–1.58 (m, 2H), 1.56–1.46 (m, 4H), 1.38–1.20 (m, 22H), 0.89 (t,
J = 7.0 Hz, 3H, CH3).
4.2.9. 10-Hydroxyoctadecanoic acid (5d) [37,41]
1H NMR (300 MHz, CDCl3) ı (ppm): 3.67–3.50 (m, 1H, CHOH),
2.34 (t, J = 7.5 Hz, 2H, CH2CO), 1.69–1.23 (m, 29H), 0.88 (t, J = 7.2 Hz,
3H, CH3); 13C NMR (100.56 MHz, CDCl3) ı (ppm): 178.8, 72.1, 37.4
(two overlapping signals), 33.9, 31.9, 29.8 (2), 29.6 (2), 29.3, 29.1,
29.0, 25.6, 25.5, 24.6, 22.7, 14.1.
4.4.2. Methyl 8-acetoxyoctadecanoate (7b) [38a,38d]
1H NMR (400 MHz, CDCl3) ı (ppm): 4.86 (quintet, 6.5 Hz, 1H),
3.68 (s, 3H, OCH3), 2.31 (t, J = 7.5 Hz, 2H), 2.04 (s, 3H, CH3CO),

44
C. Ebert et al. / Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
1.66–1.59 (m, 2H), 1.54–1.48 (m, 4H), 1.35–1.22 (m, 22H), 0.89 (t,
J = 7.2 Hz, 3H, CH3).
atom partial charges corrected and afterwards, the whole sys-
tem was submitted to a local energy minimization (including
atoms in 6 A radius from the tetrahedral carbon) and finally to a
˚
4.4.3. Methyl 9-acetoxyoctadecanoate (7c) [38a,38d,43]
1H NMR (400 MHz, CDCl3) ı (ppm): 4.86 (quintet, 6.5 Hz, 1H),
3.67 (s, 3H, OCH3), 2.30 (t, J = 7.7 Hz, 2H), 2.04 (s, 3H, CH3CO),
1.66–1.56 (m, 4H), 1.55–1.46 (m, 4H), 1.35–1.22 (m, 20H), 0.89 (t,
J = 6.7 Hz, 3H, CH3).
global energy minimization (until gradient less than 0.01 kcal/mol).
Since the fatty acid chain consists of several rotatable bonds, a
conformational search was performed with the low mode MD
conformational search tool of MOE. The lowest potential energy
conformation was chosen and then energy minimized producing
the final tetrahedral intermediate.
4.4.4. Methyl 10-acetoxyoctadecanoate (7d) [38a,38d]
1H NMR (400 MHz, CDCl3) ı (ppm): 4.84 (quintet, 6.6 Hz, 1H),
3.66 (s, 3H, OCH3), 2.29 (t, J = 7.5 Hz, 2H), 2.03 (s, 3H, CH3CO),
1.65–1.57 (m, 2H), 1.53–1.46 (m, 4H), 1.32–1.20 (m, 22H), 0.87 (t,
J = 7.2 Hz, 3H, CH3).
The analysis of the H-bond pattern was performed by means
of the in-house developed SVL tool (Scientific Vector Language in
MOE2009.10) based in molecular dynamic simulations by using
an adequate time of equilibration (energy gradient less than
0.05 kcal/mol in the last 100 ps) and productive time of 1000 ps
(time step 0.2 ps, 300 K, implicit solvent, NVT ensemble). For each
of the 5000 sampled conformations, the tool calculated the num-
ber of H-bonds involving one of the three oxygen atoms directly
connected with the tetrahedral carbon of the intermediate. Only
those H-bonds that can contribute to the stabilization of interme-
diates were taken into account (oxyanion–oxyanion hole, alcohol
moiety deriving oxygen–nitrogen of His side chain, Ser side chain
oxygen–nitrogen of His side chain). The resulting profile shows the
occurrence of poses stabilized by a defined number of H-bonds
which were observed during the production MD step.
4.4.5. Methyl
(7R)-7-{[(2R)-2-(acetyloxy)-2-phenylacetyl]oxy}octadecanoate
and methyl
(7S)-7-{[(2R)-2-(acetyloxy)-2-phenylacetyl]oxy}octadecanoate
(8a)
Within the mixture containing different relative amounts of two
diastereomers, indicated as diast1-8a and diast2-8a, it was possi-
ble to distinguish some signals of each and consequently to describe
the related spectrum. diast1-8a: 1H NMR (600 Mz, CDCl3, TMS) ı
(ppm): 7.47 (dd, J = 7.8 Hz, J = 2.2 Hz, 2H), 7.41–7.31 (m, 3H), 5.869 (s,
1H, CHPh), 4.91–4.84 (m 1H), 3.657 (s, 3H, OCH3), 2.158 (t, J = 7.6 Hz,
2H, H-2), 2.20 (s, 3H, CH3CO), 1.70–0.99 (m, 28H), 0.883 (t, J = 7.2 Hz,
3H, CH3); diast2-8a: 1H NMR (600 Mz, CDCl3, TMS) ı (ppm): 7.47
(dd, J = 7.8 Hz, J = 2.2 Hz, 2H), 7.41–7.31 (m, 3H), 5.861 (s, 1H, CHPh),
4.91–4.84 (m 1H), 3.662 (s, 3H, OCH3), 2.289 (t, J = 7.6 Hz, 2H, H-
2), 2.20 (s, 3H, CH3CO), 1.70–0.99 (m, 28H), 0.888 (t, J = 7.2 Hz, 3H,
CH3).
4.6. Modelling of tetrahedral intermediates of HSAs in the alcohol
pocket of CALB
The docking of HSA substrates was not able to generate any reli-
able pose of the alcohol moiety of HSA in the active sites of both
enzymes. Basically, all the placements algorithms implemented in
MOE gave poses with the carboxyl functionality close to the cat-
alytic machinery and, conversely, the hydroxy group out of the
deeper active site region. To obtain a structural model of the tetra-
hedral intermediates having HSA as acyl acceptor, a decarboxylated
acetyl ester derivative was docked using a pharmacophoral based
placement. The pharmacophore was designed introducing two
requirements for the docking poses: the location of the carbonyl
oxygen toward the oxyanion hole and a small hydrophobic moiety
in the acyl pocket. Using this restrains the acetylated decarboxy-
lated substrate was docked by imposing the ester group upon the
catalytic machinery (Ser-His and oxyanion hole). After the place-
ment of such decarboxylated substrates, the carboxyl group was
added to the system in order to recover the wanted functionality of
substrates. Afterwards, tetrahedral intermediates were built mak-
ing the covalent bond with the enzyme, changing the atom charges
and finally minimizing the energy until energy gradient less than
0.01 kcal/mol.
4.4.6. Methyl
(8R)-8-{[(2R)-2-(acetyloxy)-2-phenylacetyl]oxy}octadecanoate
and methyl
(8S)-8-{[(2R)-2-(acetyloxy)-2-phenylacetyl]oxy}octadecanoate
(8b)
From the mixture containing different relative amount of the
two diastereomers, indicated as diast1-8b and diast2-8b, it was
possible to distinguish some signals of each and consequently to
describe the related spectrum. diast1-8b: 1H NMR (600 Mz, CDCl3,
TMS) ı (ppm): 7.51–7.45 (m, 2H), 7.43–7.33 (m, 3H), 5.872 (s, 1H,
CHPh), 4.91–4.84 (m 1H), 3.666 (s, 3H, OCH3), 2.239 (t, J = 7.6 Hz, 2H,
H-2), 2.20 (s, 3H, CH3CO), 1.70–0.99 (m, 28H), 0.885 (t, J = 7.2 Hz, 3H,
CH3); diast2-8b: 1H NMR (600 Mz, CDCl3, TMS) ı (ppm): 7.51–7.45
(m, 2H), 7.43–7.33 (m, 3H), 5.867 (s, 1H, CHPh), 4.91–4.84 (m 1H),
3.665 (s, 3H, OCH3), 2.295 (t, J = 7.6 Hz, 2H, H-2), 2.20 (s, 3H, CH3CO),
1.70–0.99 (m, 28H), 0.890 (t, J = 7.2 Hz, 3H, CH3).
Since tetrahedral intermediates were built by forcing the sub-
strate poses with the pharmacophoral placement, a relaxation of
the system was required. Thus a molecular dynamic simulation was
performed in order to relax the tetrahedral intermediates and eval-
uate the poses (300 ps, 300 K, implicit solvent, NVT, freezing atoms
4.5. Modelling H-bonding stabilization of tetrahedral
intermediates
˚
The computational analysis was performed starting from the
crystal structures of CALB (PDB code: 1TCA) and PS (PDB code: 4LIP)
properly pretreated removing glycans and solvents. Both protein
structures were protonated (pH 7) and then relaxed by a molecular
dynamic simulation (300 K, 1000 ps, implicit solvent, NVT ensem-
ble). All the computation was performed with Molecular Operating
Environment suit (MOE2009.10 – Chemical Computing Group -
Montreal, Canada) using the Amber99 force field.
The tetrahedral intermediates for the reactions involving the
terminal carboxylic moiety, were constructed by creating the cova-
lent bond between the catalytic serine side chain and the substrate
best productive poses of docking simulations (AlphaPMI place-
ment, London dG re-scoring). The new bond was placed and the
more than 10 A far away from the tetrahedral intermediate).
4.7. 3D-QSAR
Differential Molecular Interaction Field (DMIF) has been used
as descriptors for PLS model as already published [29]. The reliable
structure of a couple of tetrahedral intermediates, the R and S enan-
tiomers of the substrate under investigation, has been analysed
calculating the molecular interaction field (MIF) with the software
GRID (Molecular Discovery Ltd., United Kingdom) using 4 probes
and taking into account a box surrounding the substrate as defined
in the cited work. The resulting pair of MIF has been used to cal-
culate differential molecular interaction field (DMIF) making the

C. Ebert et al. / Journal of Molecular Catalysis B: Enzymatic 83 (2012) 38–45
45
difference between the MIF deriving from R enantiomer and the
one from the S enantiomer. Finally the DMIF was used to make the
enantioselectivity prediction performing an external PLS prediction
with the software GOLPE (MIA s.r.l., Perugia, Italy)
[13] N. Calonghi, C. Cappadone, E. Pagnotta, G. Farruggia, F. Buontempo, C. Boga, G.L.
Brusa, M.A. Santucci, L. Masotti, Biochem. Biophys. Res. Commun. 314 (2004)
138–142.
[14] N. Calonghi, C. Cappadone, E. Pagnotta, C. Boga, C. Bertucci, J. Fiori, G. Tasco, R.
Casadio, L. Masotti, J. Lipid Res. 46 (2005) 1596–1603.
[15] R.C. Badami, K.B. Patil, Prog. Lipid Res. 19 (1981) 119–153, and references cited
therein.
Acknowledgements
[16] (a) S. El-Sharkawy, W. Yang, L. Dostal, J.P.N. Rosazza, Appl. Environ. Microbiol.
58 (1992) 2116–2122, and references cited therein;
(b) W. Yang, L. Dostal, J.P.N. Rosazza, Appl. Environ. Microbiol. 59 (1993)
281–284, and references cited therein;
(c) B.-N. Kim, S.-J. Yeom, D.-K. Oh, Biotechnol. Lett. 33 (2011) 993–997, refer-
ences cited therein.
This work has received funding from the European Commu-
nity Seventh Framework Programme under the FP7-KBBE-2008-2B
grant agreement no. 227279(IRENE project), from Alma Mater Stu-
diorum – Università di Bologna (Fundamental Oriented Research
(RFO) funds) and from Ministero dell’Istruzione, dell’Università e
della Ricerca (MIUR) (Project PRIN 20078J9L2A 005: “New fron-
tiers in the synthesis, reactions and applications of compounds
containing heteroatoms”).
[17] S. Murakami, K. Satou, T. Kijima, M. Watanabe, T. Izumi, Eur. J. Lipid Sci. Technol.
113 (2011) 450–458.
[18] R.W. Miller, D. Weisleder, R. Kleiman, R.D. Plattnera, C.R. Smith Jr., Phytochem-
istry 16 (1977) 947–951.
[19] (a) S. Bowers, H.C. Hoch, P.H. Evans, M. Katayama, Science 232 (1986) 105–106;
(b) I.D. Brodowsky, E.H. Oliw, Biochim. Biophys. Acta 1168 (1993) 68–72;
(c) U. Garscha, E.H. Oliw, Anal. Biochem. 367 (2007) 238–246, and references
cited therein.
[20] M.S.F. Lie Ken Jie, M. Khysar Pasha, M.S.K. Syed-Rahmatullah, Nat. Prod. Rep.
14 (1997) 163–189, and references cited therein.
Appendix A. Supplementary data
[21] (a) M.B. Smith, J. March, March’s Advanced Organic Chemistry: Reactions
Mechanisms, and Structure, 7th ed., John Wiley & Sons, Hoboken, 2007, pp.
1799–1805;
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.molcatb.
2012.06.017.
(b) E.M. Carreira, L. Kvaerno, Classics in Stereoselective Synthesis, Wiley-WCH,
Weiheim, 2009 (chapter 2).
[22] (a) K. Nakamura, R. Yamanaka, T. Matsuda, T. Harada, Tetrahedron: Asymmetry
14 (2003) 2659–2681;
References
(b) V. Fehring, R. Selke, Angew. Chem. Int. Ed. 37 (1998) 1827–1830;
(c) K. Goldberg, K. Schroer, S. Lütz, A. Liese, Appl. Microbiol. Biotechnol. 76
(2007) 237–248.
[1] (a) F.D. Gunstone, Fatty Acid and Lipid Chemistry, Springer Verlag, Berlin, 1996;
(b) A.J. Fulco, Prog. Lipid Res. 22 (1983) 133–160;
(c) J.L. Harwood, N.J. Russell, Lipids in Plants and Microbes, Allen and Unwin,
London, 1984.
[2] A. Sevanian (Ed.), Lipid Peroxidation in Biological Systems, The American Oil
Chemists Society, 1988.
[3] (a) G.N. McDermott, in: D. Swern (Ed.), Miscellaneous Oil and Fat Products in
Bailey’s Industrial Oil and Fat Products, John Wiley &Sons, New York, 1982;
(b) H.M. Teeter, L.E. Gast, E.W. Bell, J.C. Cowan, Ind. Eng. Chem. 45 (1953)
1777–1779;
(c) M. Farroq, A. Ramli, S. Gul, N. Muhammad, J. Appl. Chem. 11 (2011)
1381–1385.
[4] (a) E.H. Pryde, L.H.K.D. Princen, Mukherjiee New Sources of Fats and Oils, Amer-
ican oil Chemists Society, Champaign, 1981;
[23] M. Utaka, H. Watabu, A. Takeda, J. Org. Chem. 52 (1987) 4363–4368.
[24] J. Pleiss, in: U. Bornsheuer (Ed.), Enzymes in Lipid Modification, Wiley-VCH
Verlag, Weinheim, 2005, pp. 85–99.
[25] R.J. Kazlauskas, A.N.E. Weissfloch, A.T. Rappaport, L.A. Cuccia, J. Org. Chem. 56
(1991) 2656–2665.
[26] U.H. Kahlow, R.D. Schmid, J. Pleiss, Protein Sci. 10 (2001) 1942–1952.
[27] F. Haeffner, T. Norin, K. Hult, Biophys. J. 74 (1998) 1251–1262.
[28] S. Raza, L. Fransson, K. Hult, Protein Sci. 10 (2001) 329–338.
[29] (a) P. Braiuca, L.V. Knapic, C. Ferrario, L. Ebert, Gardossi, Adv. Synth. Catal. 351
(2009) 1293–1302;
(b) J. Gu, J. Liu, H. Yu, J. Mol. Catal. B: Enzym. 72 (2011) 238–247.
[30] C. Boga, S. Puggioli, M. Gherpelli, G. Farruggia, E. Pagnotta, L. Masotti, P. Neyroz,
Anal. Biochem. 335 (2004) 196–209.
[31] E. Wenkert, B.L. Buckwalter, A.A. Craveiro, E.L. Sanchez, S.S. Sharatchandra, J.
Am. Chem. Soc. 100 (1978) 1267–1273.
[32] C. Orrenius, F. Haeffner, D. Rotticci, N. Öhrner, T. Norin, K. Hult, Biocatal. Bio-
transform. 16 (1998) 1–15.
[33] (a) O. Sundholm, L.T. Kanerva, ACH Models Chem. 135 (1998) 625–640;
(b) X.-G. Li, M. Lähitie, M. Päiviö, L.T. Kanerva, Tetrahedron: Asymmetry 18
(2007) 1567–1573;
(b) M. Svensson, in: M. Kjellin, I. Johansson (Eds.), Surfactants from Renewable
Resources, vol. 1, John Wiley & Sons, Chichester, UK, 2010;
(c) U. Biermann, U. Bornscheuer, M.A.R. Meier, J.O. Metzger, H.J. Schäfer, Angew.
Chem. Int. Ed. 50 (2011) 3854–3871.
[5] (a) Z.S. Petrovic´, I. Ivetkovic´, D. Hong, X. Wan, W. Zhang, T. Abraham, J. Malsam,
J. Appl. Polym. Sci. 108 (2008) 1184–1190;
(b) T. Hamaguchi, M. Atsuhito, Plasticizer for biodegradable resin, US Patent
US20060276575 (2006).
(c) K. Lundell, T. Raijola, L.T. Kanerva, Enzyme Microb. Technol. 22 (1998)
86–93;
(d) L.T. Kanerva, O. Sundholm, Acta Chem. Scand. 47 (1993) 823–825;
[6] (a) F.C. Naughton, J. Am. Oil Chem. Soc. 51 (1974) 65–71;
(b) D. Thetford, Paint compositions, EP Patent 1,383,840 (2008);
(c) A. Marrion, A.R. Marrion, in: R. Alistair, Marrion (Eds.), Binders for Conven-
tional Coatings in the Chemistry and Physics of Coatings, vol. 1, RCS Publ., 2004,
pp. 96–150.
[7] T.F. Tadros, C. Dederen, M.C. Taelman, Parfuemerie Kosmetik 78 (1997)
30–34.
[8] (a) T.V. Drovetskaya, W.H. Yu, E.F. Diantonio, S.L. Jordan, Hair styling and con-
ditioning personal care films, US Patent US 2010/0247459 A1 (2010).;
(b) G.A. Grissett, D.M. Keenan, F.A. Macedo, D.R. Williams, Fibrous toilette arti-
cle, US Patent US 10/938,384 (2008).
[9] W. Schwab, R. Davidovich-Rikanati, E. Lewinsohn, Plant J. 54 (2008) 712–732.
[10] (a) F. Shahidi (Ed.), Nutraceutical and Specialty Lipids and Their Co-products,
CRC Press, 2006;
(b) N.E. Hughes, A.G. Marangoni, A.J. Wright, M.A. Rogers, J.W.E. Rush, Trends
Food Sci. Technol. 20 (2009) 470–480;
(c) J.F. Toro-Vazquez, J. Morales-Rueda, Food Biophys. 5 (2010) 193–202.
[11] (a) M.A. Rogers, A.J. Wright, A.G. Marangoni, Soft Matter 4 (2008) 1483–1490;
(b) M.A. Rogers, A.J. Wright, A.G. Marangoni, Curr. Opin. Colloid Interface Sci.
14 (2009) 33–42;
(e) T. Ema, Curr. Org. Chem. 8 (2004) 1009–1025;
(f) S. Tomic´, T. Kojic´-Prodic´, J. Mol. Graphics Modell. 21 (2002) 241–252.
[34] M. Skjøt, L. De Maria, R. Chatterjee, A. Svendsen, S.A. Patkar, P.R. Ostergaard, J.
Brask, Chembiochem 10 (2009) 520–527.
[35] C.-S. Chen, Y. Fujimoto, G. Girdaukas, C.J. Sih, J. Am. Chem. Soc. 104 (1982)
7294–7299.
[36] (a) B. Behrouzian, C.K. Savile, B. Dawson, P.H. Buist, J. Shanklin, J. Am. Chem.
Soc. 124 (2002) 3277–3283;
(b) P. Caruana, Bachelor Thesis, University of Bologna, Bologna (Italy), 2006.
[37] S. Bergstro˘m, G. Aulin Erdtman, B. Rolander, E. Stenhagen, S. Östling, Acta Chem.
Scand. 6 (1952) 1157–1174.
[38] (a) A.P. Tullock, J. Am. Oil Chem. Soc. 41 (1964) 833–836;
(b) A.P. Tullock, J. Am. Oil Chem. Soc. 43 (1966) 670–674;
(c) A.P. Tullock, M. Mazurek, Lipids 11 (1976) 228–234;
(d) A.P. Tullock, Org. Magn. Res. 11 (1978) 109–115.
[39] (a) R.G. Powell, C.R. Smith, C.A. Glass, I.A. Wolff, J. Org. Chem. 31 (1966) 528–533;
(b) F.D. Gunstone, A.J. Sealy, J. Chem. Soc. (1963) 5772–5778;
(c) T.H. Applewhite, R.G. Binder, W. Gaffield, J. Org. Chem. 32 (1967) 1173–1178.
[40] I.L.Z. Kuranova, Obshchei Khim 33 (1963) 2296–2301.
[41] (a) G. Knothe, M.O. Bagby, D. Weisleder, R.E. Peterson, J. Chem. Soc. Perkin Trans.
2 (1994) 1661–1670;
(c) M.A. Rogers, J.H.J. Kim, Food Res. Int. 44 (2011) 1447–1451;
(d) L. Cristofolini, M.P. Fontana, C. Boga, O. Konovalov, Langmuir 21 (2005)
11213–11219.
[12] (a) L. Masotti, E. Casali, N. Gesmundo, G. Sartor, T. Galeotti, S. Borrello, M.V.
Piretti, G. Pagliuca, Ann. N. Y. Acad. Sci. 51 (1989) 47–58;
(b) L. Masotti, E. Casali, N. Gesmundo, Mol. Aspects Med. 14 (1993) 209–215;
(c) C. Bertucci, M. Hudaib, C. Boga, N. Calonghi, C. Cappadone, L. Masotti, Rapid
Commun. Mass Spectrom. 16 (2002) 859–864.
(b) E.G. Perkins, J.R. Anfisen, J. Am. Oil Chem. Soc. 48 (1971) 556–562.
[42] I.L.Yu. Kuranova, D. Shenin, G.V.Z. Pigulevskii, Obshchei Khim 34 (1964)
3487–3493.
[43] A.P. Tullock, Can. J. Chem. 46 (1968) 3728–3729.