Production of 10(S)-hydroxy-8(E)-octadecenoic acid mono-estolides by lipases in non-aqueous media

Article abstract of DOI:10.1016/j.procbio.2012.12.006

In this study, Novozym 435, a lipase B from Candida antarctica, was used for fatty acid polymerization. For the first time, an apolar reaction media, n-hexane, was used to synthesize in vitro estolides from trans-hydroxy-fatty acids derived from the biotransformation of oleic acid by Pseudomonas aeruginosa 42A2 NCIMB 40045. We studied the effects of the substrate, the enzyme ratio, the enzyme stability and the reusability of the biocatalyst. To determine the structure of the oligomers formed, both liquid chromatography mass spectrometry and MALDI-TOF mass spectrometry, with a DHB matrix neutralized with lithium hydroxide, were used to obtain simpler mass spectra. Estolides composed of two units of (10S)-HOME were synthesized with a reaction yield of 30%. Finally, various lipases were screened, and another apolar organic solvent, iso-octane, was assayed to try to increase the reaction yield.

Full text of DOI:10.1016/j.procbio.2012.12.006

Process Biochemistry 48 (2013) 224–230
Contents lists available at SciVerse ScienceDirect
Process Biochemistry
journal homepage: www.elsevier.com/locate/procbio
Production of 10(S)-hydroxy-8(E)-octadecenoic acid mono-estolides by lipases in
non-aqueous media
I. Martin-Arjola, M. Busquetsb, A. Manresaa,∗
a Departament de Microbiologia i Parasitologia Sanitàries, Facultat de Farmàcia, Universitat de Barcelona, Av. Joan XXIII s/n, 08028 Barcelona, Spain
b Departament de Bioquímica i Biologia Molecular, Universitat de Barcelona, Av. Diagonal 643, 08028 Barcelona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 July 2012
Received in revised form
10 December 2012
Accepted 11 December 2012
Available online 20 December 2012
In this study, Novozym 435, a lipase B from Candida antarctica, was used for fatty acid polymerization.
For the first time, an apolar reaction media, n-hexane, was used to synthesize in vitro estolides from
trans-hydroxy-fatty acids derived from the biotransformation of oleic acid by Pseudomonas aeruginosa
42A2 NCIMB 40045. We studied the effects of the substrate, the enzyme ratio, the enzyme stability
and the reusability of the biocatalyst. To determine the structure of the oligomers formed, both liquid
chromatography mass spectrometry and MALDI-TOF mass spectrometry, with a DHB matrix neutralized
with lithium hydroxide, were used to obtain simpler mass spectra. Estolides composed of two units of
(10S)-HOME were synthesized with a reaction yield of 30%. Finally, various lipases were screened, and
another apolar organic solvent, iso-octane, was assayed to try to increase the reaction yield.
© 2012 Elsevier Ltd. All rights reserved.
Keywords:
Lipase
Novozym 435
trans-hydroxy-fatty acids
Estolides
MALDI-TOF
1. Introduction
The starting materials of this method were ricinoleic acid (RA),
12(R)-hydroxy-9(Z)-octadecenoic acid, or lesquerolic acid, 14(R)-
Estolides, a class of polymeric molecules containing an unsatu-
rated bond or hydroxy-fatty acids bonded to a carboxyl moiety of
another acyl group [1], have been detected in the seed oil of sev-
eral plants: Euphorbiaceae, Brasicaceae, Cruciferae, Limmanthaceae,
Compositae and, most notably, the genus Lesquerella [1–3]. Estolides
have also been found in secretions from the glandular hairs of
a caterpillar, Pieris rapae [4], and in human meibum as a wax
ester (WE) [5]. Finally, it should be noted that estolides were also
detected in cultures of Pseudomonas aeruginosa 42A2 when culti-
vated with oleic acid [6].
There is great interest in producing estolides and their deriva-
tives as biodegradable lubricants, functional fluids, cosmetics, inks
or coatings [7]. To this end, the first attempt to produce estolides
relied on the polymerization of ricinoleic acid from castor oil [8] or
oleic acid in a reaction that required a temperature of 250 ◦C under
pressure [9]. Currently, estolides are produced chemically from
vegetable oils (soybean, meadow foam, lesquerella, castor, coconut
or palm kernel oils) composed of unsaturated fatty acids of ten or
more carbon atoms [9–11]. To overcome these extreme conditions
and to prevent side effects due to the use of an inorganic catalyst, a
biocatalytic method with random lipases from Candida rugosa [12],
Candida antarctica or Rhizomucor miehei [13,14] was developed.
hydroxy-11(Z)-icosenoic acid [1,15].
Although there are a large number of reports about the bio-
transformation of lipids with lipases, there are few studies on the
polymerization of hydroxy-fatty acids by lipases. In 1989, Yam-
aguchi et al. were the first to report the esterification of castor oil in
a two-step enzymatic reaction to obtain estolides with a high yield
[16]. Later, they reported the importance of controlling the amount
of water present in the reaction media and showed the advan-
tage of immobilizing the lipases; the reaction yield was improved,
and the product obtained reached acid values (AV) of 40 [17–19].
With estolides from ricinoleic acid applications arising, different
strategies have been tested: (i) improving the immobilizing carri-
ers [19,20]; (ii) encapsulating lipases in a reverse micelle [1]; and
(iii) from 1997 to 2009, optimized reaction conditions in a biore-
actor to increase the quality and quantity of estolides production
were reported [12,21–23]. More recently, organic apolar solvents
have started to be used as part of the reaction media [24] and to
introduce new applications of ricinoleic acid estolides in the food
industry [12]. Likewise, molecular sieves have been employed to
adsorb water released during polycondensation in order to shift
thermodynamic balance [25]. Several other methods, such as air
drying [12], nitrogen bubbling, vacuum pressure, or a combination
of these two, have also been employed [26].
Little has been documented about the polymerization of
trans-hydroxy-acids with specific lipases toward trans-monomers.
The studies by Borgdorf and Warwel, in which thirty-nine
∗
Corresponding author.
E-mail address: amanresa@ub.edu (A. Manresa).
1359-5113/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.procbio.2012.12.006

I. Martin-Arjol et al. / Process Biochemistry 48 (2013) 224–230
225
commercial lipases were classified according to a ratio competitive
factor (RCF), are very helpful in determining substrate selectivity
toward isomeric 9-octadecenoic acids in n-hexane [27].
Quantitative analysis of OA, (10S)-HPOME, (10S)-HOME and (7S,10S)-DiHOME
was carried out by liquid chromatography in a Shimadzu LC-9A Chromatograph
(Kyoto, Japan). Samples were injected into the HPLC with a Sedex 55 light-scattering
detector (Sedere, Alfortville Cedex, France) equipped with a Tracer Excel 120 C8 col-
umn (5 ␮m, 150 mm × 4.6 mm) (Teknokroma, Sant Cugat del Vallès, Spain). Optimal
separation was achieved with an elution gradient using A, acetonitrile (Fischer Sci-
entific, Madrid, Spain) (0.1%, v/v acetic acid), and B, water (0.1%, v/v acetic acid),
at a flow rate of 1 ml min−1. The gradient (time (min), %B) used was as follows: (0,
70), (10, 0), (15, 0), (20, 70), and (25, 70). The injection volume was 20 ␮l. A known
homemade standard of each hydroxy-fatty acid and substrate was used to identify
the retention times and to quantify the samples. Cell growth was calculated as dry
cell weight. The biomass of the samples was placed in an oven at 100 ◦C for 24 h. The
nitrate concentration was determined using QUANTOFIX® nitrate/nitrite test strips
It is a difficult to both detect and characterize estolides. Gas
chromatography coupled with mass spectrometry has proven use-
less due to the high boiling point of estolides [28–30]. In recent
years, matrix-assisted laser desorption/ionization time-of-flight
mass spectrometry (MALDI-TOF-MS) has emerged as a powerful
technique to analyze synthetic polymers. Cvacka and co-workers
analyzed WE using the MALDI-TOF-MS technique with a matrix
of lithium salt of 2,5-dihydroxybenzoic acid (DBH) to increase the
intensity of the signal for obtaining a good reproducibility [31].
They were able to characterize natural samples of WE with great
accuracy.
(Macherey-Nagel, Düren, Germany). All concentrations are expressed in g l−1
.
2.3. Purification of trans-hydroxy-fatty acids
At the end of cultivation, the culture was centrifuged (14,700 × g for 30 min
at 4 ◦C), the supernatant was acidified to pH 2.0 with 37% HCl (Panreac, Castellar
del Vallès, Spain), and two extractions were performed with a half volume of ethyl
acetate (Carlo-Erba Reagents-SdS, Sabadell, Spain). The organic phase was dried over
an anhydrous sodium sulfate filter (Panreac, Castellar del Vallès, Spain), and the
solvent of the combined extracts was evaporated with a rotary evaporator (Bücchi,
Postfach, Switzerland), resulting in a yellow oil. Organic extracts were kept in vials
at 4 ◦C under nitrogen to prevent further oxidation. The (10S)-HOME and (7S,10S)-
DiHOME were purified by flash-chromatography in a glass column (50 cm long and
with 3 cm inner diameter) filled with silica gel 60 (0.040–0.063 mm, Merck, Madrid,
Spain). The mobile phase used was formed by n-hexane:diethyl ether:methanol
(80:20:7, v:v:v) and a stream of nitrogen (Carburos Metálicos, Barcelona, Spain)
The primary goal of this research is to develop new bio-based
materials from agroindustrial wastes. To this end, oleic acid, which
is present to a high degree in residual oily wastes, was used as
a substrate for the production of 10(S)-hydroxy-8(E)-octadecenoic
and 7,10(S,S)-dihydroxy-8(E)-octadecenoic acids ((10S)-HOME and
(7S,10S)-DiHOME, respectively) by P. aeruginosa 42A2. Next, the
monomer (10S)-HOME was used to synthesize estolides in vitro
with Novozym 435. MALDI-TOF-MS and LC–MS techniques were
used to determine the estolides’ structure. Finally, lipases were
screened in an organic medium to improve the reaction yield of
the catalytic synthesis of estolides.
was applied to obtain the purified products. The purified products were kept at 4 ◦
under nitrogen, as stated above.
C
2. Materials and methods
2.4. Product yields and productivity of monomers
2.1. Materials
The whole specific product yield (YP/X (Eq. (1))) and specific (10S)-HOME and
(7S,10S)-DiHOME yields (Y(10S)-HOME/X and Y(7S,10S)-HOME/X (Eq. (2))), were calculated
P. aeruginosa 42A2 NCIMB 40045 was maintained on TSA (Trypticase soy agar;
Difco, Franklin lakes, IL, USA) slants at 30 ◦C and after an incubation period of 24 h
was kept at 4 ◦C and subcultured every two weeks. The strain was preserved frozen
(−80 ◦C) in cryobilles (AES CHEMUNEX S.A., Terrassa, Spain). A lipase from C. rugosa
(1104 units/mg solid) was purchased from Sigma–Aldrich (Madrid, Spain). Novozym
435, immobilized lipase B from C. antarctica (10,000 propyl laureate units/g solid),
Lipozyme RM IM (an immobilized lipase from R. miehei, 275 interestification units/g
solid) and Lipozyme TL IM (a lipase from Thermomyces lanuginosus, 250 interestifi-
cation units/g solid) were generous gifts from Novozym A/S (Bagsvaerd, Denmark).
Lipase A from Rhizopus oryzae was donated by Prof. Francisco Valero (Chem Eng
Dept UAB, Cerdanyola del Vallès, Spain). Analytical grade n-hexane, methanol and
diethyl ether were purchased from Carlo-Erba Reagents-SdS (Sabadell, Spain). Ana-
lytical grade iso-octane for spectroscopy UV–IR and potassium hydroxide (KOH) 85%
were purchased from Panreac (Castellar del Vallès, Spain). Ricinoleic acid (RA) 80%
was supplied by Fluka (Madrid, Spain) and oleic acid (OA) technical grade 90% was
purchased from Sigma–Aldrich (Madrid, Spain).
over the entire process as the productivity of each hydroxy-fatty acid, q(10S)-HOME
q(7S,10S)-DiHOME, instead of the OA consumption rate, qOA, which was calculated over
the first 15 h of cultivation (Eq. (3)).
,
ꢀC((10S)-HOME) + ꢀC(7S, 10S)-DiHOME)
YP/X (g
=
(1)
(2)
−1
g
)
ꢀC(X)
ꢀC(10S)-HOME
ꢀC(X)
Y(10S)-HOME/X (g
=
;
−1
g
)
ꢀC(7S, 10S)-DiHOME
=
ꢀC(X)
Y(7S,10S)-DiHOME/X (g
−1
g
)
ꢀC((7S, 10S)-DiHOME)
ꢀC(OA)
ꢀt
q(7S,10S)-DiHOME (g l−1 h−1) =
;
qOA (g l−1 h−1) =
ꢀt
(3)
2.2. Production and quantification of monomers
2.5. Measurement of the reaction yield of estolides synthesis
P. aeruginosa 42A2 NCIMB 40045 was cultured in the following saline medium
(in g l−1): CaCl2 (0.01), NaNO3 (3.5), K2HPO4 (2.0), KH2PO4 (1.0), KCl (0.1),
MgSO4·7H2O (0.5), FeSO4·7H2O (0.012) and 0.05 ml l−1 trace elements solution. The
trace elements solution was as follows (in mg 100 ml−1): H3BO3 (148), CuSO4·5H2O
(196), MnSO4·H2O (154), Na2MoO4·2H2O (15) and ZnSO4·7H2O (307). This medium
was supplemented with OA at 2% (v/v).
The reactions were monitored by titration to determine the acid value (AV)
[23] of the samples. After evaporating the organic solvent, a 30 mg aliquot of the
reaction mixture was titrated with 0.05 M KOH using phenolphthalein as indicator.
All samples were analyzed in triplicate. The AV and the yield of the reaction were
calculated as follows (Eq. (4)):
The bioreactor was inoculated with 200 ml of a 24 h culture in a saline medium
containing 20 g l−1 OA as a carbon source. The inoculum-culture was carried out
on an orbital shaker for 18 h at 150 rpm rotational speed and 30 ◦C. Cells were har-
vested by centrifugation (14,700 × g for 30 min at 4 ◦C) and resuspended with NaCl
0.9% (w/v) at 2% (v/v) to an optical density of 2.0 at 540 nm prior to inoculation
into the bioreactor. The cultures were cultivated at a working volume of 2 l in a 3-
L bench top bioreactor (Biostat B. Braun Biotech International GmbH, Melsungen,
Germany) using the mineral salts medium described above. During the culture, dis-
solved oxygen was monitored continuously with an O2 electrode (Ingold 12/200 B.
Braun Biotech. International GmbH, Melsungen, Germany) and maintained at 30%
oxygen saturation by automatic cascade control of the stirrer speed, 500–700 rpm,
with an air flow between 2.5 and 7.5 l min−1. The air flow was enriched with oxy-
gen (Carburos Metálicos, Spain) when needed. The temperature was measured at
30 ◦C by a Pt-100/200-4 temperature sensor (B. Braun Biotech. International GmbH,
Melsungen, Germany). The pH was automatically kept at 7.0 using 2 M HCl and 2 M
NaOH solutions. The data were recorded using an external computer connected to
the control unit of the bioreactor. The software used was MFCS/win 2.0 (B. Braun
Biotech International, Sartorius, Mesulgen, Germany). The culture was maintained
until all 10(S)-hydroperoxy-8(E)-octadecenoic acid ((10S)-HPOME) was converted
to (10S)-HOME and (7S,10S)-DiHOME.
AVsubstrate − AVproduct
ꢁ (%) =
· 100
(4)
AVsubstrate
2.6. Effect of enzyme concentration
Different quantities of Novozym 435 (0.3–1.5 g) were assayed to obtain the opti-
mal amount of enzyme for a reaction with 0.6 g of (10S)-HOME in 20 ml of n-hexane
in a 100 ml Erlenmeyer flask for 48 h at 50 ◦C. A rotary evaporator system was used to
achieve an efficient degree of contact between the enzyme and the substrate and to
maintain the required temperature. All reactions were carried out at atmospheric
pressure. The reaction extension was calculated using Eq. (4), as shown above. A
control was assayed to confirm that this reaction does not occur spontaneously.
2.7. Effect of substrate/enzyme ratio with Novozym 435
An aliquot of 0.5 g of Novozym 435 was used to study polymerization with dif-
ferent amounts of (10S)-HOME (0.25–1.0 g). The enzymatic reaction was performed
under the same conditions described above and included a control. The yield of the
reaction was calculated with Eq. (4).

226
I. Martin-Arjol et al. / Process Biochemistry 48 (2013) 224–230
Oleic Acid
NO3-
(7S)-HOME
(7S,10S)-DiHOME
2.8. Reusability of the Novozym 435 biocatalyst
DryCell Weight
The biocatalytic reaction was performed with 0.5 g of organic substrate (10S)-
HOME, which was added to 0.5 g of enzyme. The reaction conditions were the same
as described above. After each cycle, the enzyme was removed from the reaction
medium by filtration and then rinsed with n-hexane several times. Finally, the sol-
vent was evaporated under a stream of air at room temperature. The product was
analyzed, and the stability of the enzyme was determined as the reaction yield.
24
20
16
12
8
7
6
5
4
3
2
1
0
2.9. Screening of other lipases for estolides formation
An aliquot of 100 mg of C. rugosa lipase, lipase A from R. oryzae, Lipozyme RM
IM, Lipozyme TL IM and Novozym 435 were assayed with 0.6 g of (10S)-HOME in
20 ml of n-hexane or iso-octane in a 100 ml Erlenmeyer flask for 48 h at 50 ◦C. A
rotary evaporator was used to attain a high mixing degree and to control the nec-
essary temperature. The polymerization reactions were carried out at atmospheric
pressure, and the reaction extension was calculated with Eq. (4).
4
0
0
5
10
15
20
25
30
Time (h)
Fig. 1. Time course of growth and product formation of P. aeruginosa 42A2 when
cultivated in a mineral medium in a 3-L B. Braun Biotech reactor (see text for condi-
tions). OA: (ꢀ); (10S)-HOME: (ꢁ); (7S,10S)-DiHOME: (ꢂ); NO3−: (᭹); dry cell weight:
(*).
2.10. Liquid chromatography coupled to mass spectrometry (LC–MS)
Chromatographic separation was carried out in a Perkin-Elmer (Whaltham, MA,
USA) Series 200 liquid chromatographer coupled to a PE SCIEX API 150 EX single-
quadrupole mass spectrometer (Applied Biosystems, Carlsbad, CA, USA). A Tracer
Kromasil 100 C8 column (250 mm × 46 mm, 5 ␮m) (Teknokroma, Sant Cugat del
Vallès, Spain) was used. Separation was achieved with a gradient elution using A:
acetonitrile (0.1%, v/v acetic acid); B: acetone (0.1%, v/v acetic acid) at a flow rate of
0.5 ml min−1. Gradient (time (min), %B): (0, 35), (25, 100), (30, 100), (35, 35), (40, 35).
All reported data were acquired with an APCI ionization source in negative mode
using the following parameters: vaporizer temperature of 400 ◦C, 3 mA nebulizer
current, −25 V declustering potential, −110 V focusing potential, −10 V entrance
potential and nitrogen as a nebulizer and curtain gas with 10 and 12 arbitrary units.
The scan data were obtained by scanning from m/z 100–2000 amu. For the analysis of
estolides from ricinoleic acid, the selected ion monitoring (SIM) technique was used
to detect high molecular mass oligomers with a better signal. The ions selected were
those corresponding to the different oligomers synthesized. All data were recorded
using Analyst Software v. 1.4.2 (Applied Biosystems, Carlsbad, CA, USA).
3.2. Novozym 435 for trans-hydroxy-fatty acid isomer
esterification: enzyme amount and substrate/enzyme ratio
In terms of selectivity to a cis/trans configuration of the double
bond in 9-octadecenoic acid isomers, the lipases were classified
according to a ratio competitive factor (RCF) [27]. This factor
describes the selectivity of one single lipase toward two substrates
with the same leaving group and to two different acyl groups.
Thus, in this study, Novozym 435, with an RCF of 0.7, was chosen
because it is slightly more active with trans-9-octadecenoic iso-
mers, it exhibits versatility, and its optimal temperature of enzyme
activity is 40–60 ◦C. To generate compatible new compounds in cos-
metic or food applications, n-hexane with a log P of 3.5 [33] was
chosen. Log P is a measurement of solvent polarity. Along with the
partition coefficient and the enzyme/solvent interaction, log P may
determine biocatalytic activity; this activity is low in solvents at
log P < 2, moderate at log P between 2 and 4 and high in a polar
solvent with log P > 4 [34].
Different amounts of enzyme were tested to establish the opti-
mal amount of biocatalyst for 0.6 g of (10S)-HOME. The results are
given in Fig. 2. It could be observed that the reaction yield increased
with the amount of enzyme. At values near 0.5 g of enzyme, this
yield was 35%, and it increased slightly to 42% at a substrate/enzyme
ratio of 1.2 g g−1. Thus, the thermodynamic limit for yield was very
similar to this latter value. There are no data in the literature about
polymerization with trans monomers, although this value is lower
than the 58% obtained by Horchani et al. when using ricinoleic acid
to produce estolides in n-hexane for 55 h at 55 ◦C [25]. Bódalo et al.
obtained a 72% reaction yield when C. rugosa was immobilized in
Lewatit Monoplus MP64, an anion exchange resin, for 150 h at 40 ◦C
in a solvent-free system [23].
2.11. MALDI-time-of-flight mass spectrometry (TOF-MS) structural determination
MALDI-TOF-MS experiments were performed on a 4800 Plus MALDI-TOF/TOF
Analyzer (ABSciex – 2010, Framingham, MA, USA). The samples were analyzed in
MS Reflector mode using positive ion detection. Desorption and ionization were
achieved with an Nd-YAG laser (355 nm, 3–7 ns pulse, 200 Hz). The data were
recorded using 4000 Series Explorer software (ABSciex – 2010, Framingham, MA,
USA). A 2,5-dihydroxybenzoic acid matrix was neutralized with lithium hydroxide
(Sigma–Aldrich, Madrid, Spain), as Cvacka and co-workers established [31]. Lithium
salt matrix, 10 mg ml−1, was dissolved in a mixture of acetone:trichloromethane
(2:1, v:v). An aliquot of 1 ␮l of the matrix solution was spotted on the target plate
until the organic solvent evaporated completely. Then, 1 ␮l of the samples diluted
in 2.5 mg ml−1 trichloromethane was deposited over the matrix spot and allowed
to dry before analysis.
3. Results and discussion
3.1. Production of trans-hydroxy-fatty acids by submerged
fermentation
P. aeruginosa 42A2 NCIMB 40045 was isolated from an olive
oil-contaminated site. This strain catalyzes the conversion of oleic
acid (OA) into new derivatives of the trans-configuration: 10(S)-
hydroxy-8(E)-octadecenoic acid ((10S)-HOME)) and 7,10(S,S)-
dihydroxy-8(E)-octadecenoic acid ((7S,10S)-DiHOME)) [32]. As
Fig. 1 shows, the maximum production of (10S)-HOME and
(7S,10S)-DiHOME was achieved at the end of the incubation
(8.6 and 9.2 g l−1, respectively), i.e., 89% of substrate conver-
sion. The entire specific product yield, YP/X, was 2.62 g g−1, and
the OA consumption rate, qOA, was 1.267 g l−1 h−1. The pro-
ductivity of (10S)-HOME, q(10S)-HOME, was 0.237 g l−1 h−1 with a
cell yield, Y(10S)-HOME/X, of 1.23 g g−1. The (7S,10S)-DiHOME pro-
ductivity, q(7S,10S)-DiHOME, was 0.267 g l−1 h−1, and its cell yield,
Y(7S,10S)-DiHOME/X, was 1.39 g g−1. After organic extraction (10S)-
HOME was purified up to 91% purity, and (7S,10S)-DiHOME was
purified to 96%. Table 1 summarizes the calculations of trans-
hydroxy-fatty acid production from OA by P. aeruginosa 42A2 in
the bioreactor.
Various substrate/enzyme ratios were tested to establish the
optimal concentration of substrate to be used. As Fig. 3 shows, the
yield increased up to a ratio of 1, after which the yield remained
Table 1
Experimental data of the production process of (10S)-HOME. i: OA, (10S)-HOME,
(7S,10S)-DiHOME and products.
Compound
Yi/X (g g−1
)
qi (g l−1 h−1
)
Purity (%)
Purification
recovery (%)
OA
3.42
1.23
1.39
2.62
1.267
0.237
0.267
0.504
–
91
96
–
–
43
29
–
(10S)-HOME
(7S,10S)-DiHOME
Products
OA: oleic acid; (10S)-HOME: 10(S)-hydroxy-8(E)-octadecenoic acid; (7S,10S)-
DiHOME: 7,10(S,S)-dihydroxy-8(E)-octadecenoic acid; products: (10S)-HOME and
(7S,10S)-DiHOME. Data are the average of three replicates.

I. Martin-Arjol et al. / Process Biochemistry 48 (2013) 224–230
227
(%)
50
40
30
20
10
0
Novozyme 435
40
30
20
10
0
0.70
0.65
0.60
0.55
0.50
1
2
3
4
5
6
7
8
9
10
0
0.5
1
1.5
Cycle
Novozym 435 (g)
Fig. 4. Yield of the polymerization and enzyme weight at each cycle. In all exper-
iments 0.5 g of Novozyme 435 and 0.5 g of (10S)-HOME were dissolved in 20 ml of
n-hexane during 48 h at 50 ◦C. Yield is calculated as percentage of the reduction of
AV. Novozyme 435: (ꢁ); bars: yield.
Fig. 2. Yield of the polymerization with different amounts of enzyme, Novozyme
435, in 0.6 g of (10S)-HOME at 50 ◦C during 48 h in n-hexane, 20 ml. Yield is calculated
as percentage of the reduction of AV.
constant at 30%, indicating that an excess substrate concentration
is not effective in enhancing the AV decrease. A similar ratio was
found by Langone and co-workers after the conversion of oleic
acid/methyl ricinoleate to estolides using Novozym 435 as a bio-
catalyst in a solvent-free system for 48 h at 80 ◦C [34].
n-hexane for 1 h, indicating a negligible effect on enzyme stability.
This indicates that using an appropriate apolar organic solvent may
help maintain enzyme stability and solubilizing the substrate may
help so as not to affect enzyme activity.
3.4. Structural determination of (10S)-HOME estolides
3.3. Effect of enzyme stability
A gel-permeation analysis of the polymers has certain dif-
ficulties that we tried to overcome, such as the difficulty of
separating different degrees of polymerization properly, especially
when the molecular mass of the oligomers analyzed does not
differ enough. At this juncture, more than one gel permeation
column is needed to achieve a good separation, which increases
the analysis time and makes this technique tedious and expen-
sive. Alternatively, Bayer et al. introduced coordination-ion-spray
mass-spectrometry (CIS-MS), in which non-polar compounds or
substances with weakly basic or acidic groups are detected, but
with poor sensitivity [36]. MALDI-TOF is ideally suited for poly-
mer analysis because of the simplicity of the mass spectra [37],
which show mainly single-charged quasi-molecular ions with lit-
tle fragmentation when cationization salts are used. To this end,
lipase-formed estolides from (10S)-HOME were analyzed using
ricinoleic acid estolides that were enzymatically produced as a con-
trol to confirm the validity of the following structural techniques:
liquid chromatography–mass spectrometry and MALDI-time-of-
flight mass spectrometry.
The use of enzymes or immobilized enzymes for repeated use
may help decrease product cost and make the enzymatic process
economically viable. The ability of lipase B of C. antarctica, Novozym
435, to retain its stability during recycling by using fatty acids as
substrate was studied by several researchers [34,35].
Fig. 4 shows the profile of the polymerization yield during dif-
ferent cycles. As observed, the synthetic stability of the enzyme
decreased throughout the cycles assayed. The yield decreased by
53.3% after ten cycles, from 30.4% to 18.5%, whereas the enzyme
weight increased. Although the biocatalyst was rinsed several times
with n-hexane, a portion of the substrate remained adsorbed on
the support due to the poor solubility of the (10S)-HOME in n-
hexane. Langone and co-workers observed a similar reduction in
enzyme stability (55%) in the production of estolides from oleic acid
and methyl ricinoleate with Novozym 435 in four batches [34]. In
contrast, Radzi et al. [35] observed great synthetic stability even
after nine cycles (91.9%) during the production of oleyl oleate in
40
30
20
10
0
3.4.1. Liquid chromatography–mass spectrometry (LC–MS)
structural determination
Estolides from RA were enzymatically produced using the
method described by Bódalo et al. [22]. Thus, RA estolides, with
an AV of 68 mgKOH g−sa1mple, were used to confirm the validity of
the LC–MS techniques applied in order to determine the struc-
ture of the apolar polymers as polyesters of C18-hydroxyl-fatty
acids. In addition, selected ion monitoring (SIM) was used to detect
high molecular mass oligomers with greater intensity. Six ions
were selected, m/z 298, 578, 858, 1138, 1419 and 1699, that corre-
spond to RA and its oligomers [2RA-H2O], [3RA-2H2O], [4RA-3H2O],
[5RA-4H2O] and [6RA-5H2O], respectively. The six-extracted ion-
superimposed chromatogram (Fig. S1, Supporting Information)
gives the retention times at which various selected ions are
detected. As can be observed, the ion with m/z 298 appears at
7.28 min peak is unreacted RA; and at the peaks (min) 12.07, 18.70,
24.29, 28.22 and 30.94, an ion with m/z 298 appears as a marker of
the cleavage of one of the ester bonds in the oligomers synthesized.
However, ions with m/z 578 (red), 858 (green) and 1138 (gray) are
0
0.5
1
1.5
2
(10S)-HOME/Novozyme 435 (g g-1)
Fig. 3. Yield of the polymerization with different amounts of (10S)-
HOME/Novozyme 435 ratio. 0.5 g of Novozyme 435 were used in all experiments
in 20 ml of n-hexane as organic solvent at 50 ◦C during 48 h. Yield is calculated as
percentage of the reduction of AV.

228
I. Martin-Arjol et al. / Process Biochemistry 48 (2013) 224–230
Table 2
Reaction yield, ꢁ (%), in the lipase screening with different apolar solvents.
Lipase
RCFa
n-hexane
Iso-octane
Specificity
Novozyme 435
Lipozyme RM IM
Candida rugosa
Rhizopus oryzae lip A
Lipozyme TL IM
0.7
1.3
2.9
11.7
26.4
16.8
16.0
23.6
12.5
13.8
32.7
Non specific
sn-1,3
Non specific
sn-1,3
3.7–4.1b 15.1
–
36.4
sn-1,3
Data are the average of two replicates.
a
Ratio competitive factor (1/˛)cis/(1/˛)trans
From Rhizopus arrhizus and Rhizopus delemar (Borgdorf and Warwel [27]).
.
b
of the 7.25 min peak is shown in Fig. 5b, and an ion with m/z 297.3
corresponding to the [M−H]− ion of (10S)-HOME can be detected.
Likewise, the mass spectrum of the 11.78 min peak shows two main
ions (Fig. 5c), m/z 577.8 and 297.5. The first ion indicates the pres-
ence of an oligomer of two units of (10S)-HOME, [2M-H2O-H]−,
and the second one represents the cleavage of the ester bond of the
same compound.
3.4.2. MALDI-time-of-flight mass spectrometry (TOF-MS)
structural determination
Selecting an appropriate MALDI matrix, cationization salts and
sample preparation techniques are critical success factors for
obtaining a reliable mass spectrum from which to infer structural
information [31]. A 2,5-dihydroxybenzoic acid (DHB) matrix, neu-
tralized with lithium hydroxide, was used to analyze RA estolides
using the method developed by Cvacka and co-workers in order
to analyze wax esters, esters from a fatty acid and a fatty alcohol.
As can be observed (Fig. S2, Supporting Information), a relatively
low-intensity peak of m/z 305.19 corresponded to the lithium
molecular adduct of RA [RA+Li7]+; a peak of m/z 287.19 appeared
due to the loss of water from the RA lithium adduct. Three peaks
in the central part of the spectra (585.53, 865.65 and 1145.87 m/z)
stand out, and these peaks correspond to molecular adducts with
lithium: [2RA-H2O+Li7]+, [3RA-2H2O+Li7]+ and [4RA-3H2O+Li7]+.
Finally, a very low-intensity peak, which coincides with the mass
of the [5RA-4H2O+Li7]+ oligomer, appears in the high-mass field
of the spectra, at m/z 1426. In contrast to the findings of Hayes
and Kelly [38], who used a trans-3-indoleacrylic acid matrix with a
sodium chloride solution to analyze polyhydric alcohol-poly (rici-
noleic acid) species, lithium DHB salt matrix could be used to
detect quasi-molecular ions with a molecular weight lower than
500 Da. Other matrices, DHB and ␣-cyano-4-hydroxycinnamic acid,
without cationization salts were tested to detect ricinoleic acid
estolides, but highly fragmented mass spectra were obtained, mak-
ing it tedious and difficult to identify the quasi-molecular ions.
When a sample of (10S)-HOME polymerized with Novozym
435 in n-hexane was analyzed with an AV of 127.8 mgKOH g−1
sample
(Fig. 6), two groups of peaks were found. In the first group, the
lithium molecular adduct of the monomer [M+Li7]+ with m/z 305.2
was observed. In the second group, a peak of m/z 585.4 was
observed, which corresponds to the lithium adduct of the oligomer
formed by two units of (10S)-HOME [2M-H2O+Li7]+, thus con-
firming the results found by liquid chromatography.
Fig. 5. (a) Full scan chromatogram of (10S)-HOME polymerized with Novozyme
435 at 50 ◦C during 48 h; AV of the sample of 127.8 mgKOH g−sa1mple. (b) Mass spectra
of the peak of 7.25 min retention time: (10S)-HOME. (c) Mass spectra of the peak of
11.78 min retention time: dimmer of (10S)-HOME.
3.5. Screening of other lipases for estolides formation
only observed at retention times of 12.07, 18.70 and 24.29 min,
respectively. Finally, ions with m/z 1419 and 1699 (not in the fig-
ure) are only detected at retention times of 28.22 and 30.94 min,
respectively.
Finally, some other lipases were tested to improve the conver-
sion rate of estolide production. An aliquot of 0.6 g of (10S)-HOME
was assayed with 100 mg of different lipases in both apolar organic
solvents, n-hexane and iso-octane (log P = 4.5 [33]), for 48 h at 50 ◦C.
Table 2 summarizes the reaction yields, and the lipases are listed
according to their ratio competitive factors (RCF). It is notable that
there is no relationship between the RCF and the yield of the reac-
tion when the reaction proceeds in an aqueous-free solvent. It was
A
sample of (10S)-HOME polymerized in n-hexane with
Novozym 435 was analyzed, with an AV of 127.8 mgKOH g−1
.
sample
As seen from Fig. 5a, two noticeable peaks at 7.25 and 11.78 min
were detected in the full scan chromatogram. The mass spectrum

I. Martin-Arjol et al. / Process Biochemistry 48 (2013) 224–230
229
Fig. 6. MALDI-TOF mass spectra of (10S)-HOME polymerized with Novozyme 435 at 50 ◦C during 48 h using a matrix of DHB neutralized with LiOH; AV of the sample of
127.8 mgKOH g−1
.
sample
expected that higher reaction yields would be found with Novozym
435 (11.7%) due to its low RCF (0.7), but the observed values were
similar to those of C. rugosa and R. oryzae lipases (16.8% and 15.1%,
respectively), which both have an RCF higher than 1. On the other
hand, Lipozyme RM IM and Lipozyme TL IM had RCF values of 26.4%
and 36.4%, respectively. Although the specificity of the assayed
lipases was listed with the aim of establishing a correlation with
the yield observed, it seems that the non-specific lipases (Novozym
435 and C. rugosa) had lower reaction yields than did the sn-1,3-
specific lipases (Lipozyme RM IM, Lipozyme TL IM and R. oryzae lip
A) with the exception of lip A from R. oryzae. However, Bódalo found
that 1,3-selective lipases are unable to attack secondary alcohols
[22]; perhaps, the configuration of the double bond and its relative
position to the secondary alcohol had an important effect.
In addition, iso-octane, another apolar organic solvent with a
higher log P value was tested (4.5). Unlike n-hexane, iso-octane is
not a compatible solvent in food applications. However, estolides
produced with this apolar solvent can be used as lubricants in cos-
metics, inks or coatings [7]. Iso-octane was selected to increase the
solubility of the substrate (10S)-HOME and, thus, the reaction yield,
but no significant differences were observed. Given these results, a
future study of the production of estolides with Lipozyme RM IM
or Lipozyme TL IM and an apolar organic solvent that enhances the
solubility of (10S)-HOME is needed. In such a study, there would be
an increase in the polymerization conversion to test the physical
properties of this new family of trans-estolides.
were obtained and the identity of the quasi-molecular ions of the
products became clear. Estolides of two units of (10S)-HOME were
detected. Because the results show no high conversion yields, a new
array of reaction media with diverse apolar solvents and lipases is
possible.
Acknowledgments
This work was supported by the Ministerio de Economía y Com-
petitividad (CICYT, project CTQ2010-21183-C02-01), Spain, and
by the IV Pla de Recerca de Catalunya (Generalitat de Catalunya)
grant 2009SGR819. I. Martin-Arjol is a grateful recipient of an
APIF-fellowship from the University of Barcelona. We also thank
Novozyms for kindly providing the lipase samples and Dr. F. Valero
(Dept. Ingenyeria Química, Universitat Autònoma Barcelona, Spain)
for the R. oryzae lipase samples, Dr. I. Ferández-Vidal and Dr. A.
Adeva from the Centres Científics i Tecnològics (CCiT) of the Uni-
versity of Barcelona who performed the spectrometric analysis of
the samples.
Appendix A. Nomenclature
(10S)-HOME 10(S)-hydroxy-8(E)-octadecenoic acid
(7S,10S)-DiHOME 7,10(S,S)-dihydroxy-8(E)-octadecenoic acid
amu
AV
atomic mass units [Da]
acid value [mgKOH g−1
]
sample
4. Conclusions
C((10S)-HOME) concentration of (10S)-HOME [g l−1
]
C((7S,10S)-DiHOME) concentration of (7S,10S)-DiHOME [g l−1
]
For the first time, estolides from a trans-hydroxy fatty acid (10S)-
HOME, have been synthesized in vitro using Novozym 435, lipase
B of C. antarctica, in an apolar organic medium and n-hexane. The
reaction media were optimized to achieve a 30% reaction yield, at
which point two structural techniques were used to determine the
nature of the synthesized products: MALDI-TOF-MS and LC–MS.
By adapting the analysis conditions, especially using a lithium-
DHB-salt matrix for the MALDI experiments, simpler mass spectra
C(X)
CIS
concentration of biomass as dry cell weight [g l−1
coordination-ion-spray
]
CKOH
Da
concentration of potassium hydroxide [M]
Daltons
DHB
LC
2,5-dihydroxybenzoic acid
liquid chromatography
log P
M
partition coefficient [–]
(10S)-HOME

230
I. Martin-Arjol et al. / Process Biochemistry 48 (2013) 224–230
MALDI matrix-assisted laser desorption/ionization
[13] Awang R, Basri M, Ahmad S, Salleh AB. Enzymatic esterification of dihydroxys-
tearic acid. JAOCS 2000;77:609–12.
[14] Gosh M, Bhattacharyya DK. Lipase-catalyzed synthesis of hydroxy stearates and
their properties. JAOCS 1998;75:1057–9.
[15] Hayes DG. The catalytic activity of lipases toward hydroxy fatty acids – a review.
JAOCS 1996;73:543–9.
[16] Yamaguchi C, Akita M, Asaoka S, Osada F. Enzymic manufacture of castor-
oil fatty acid estolides. Japanese Kokai Tokkyo Koho 1989, JP 01016591 A
19890120.
[17] Yoshida Y, Kawase K, Yamaguchi C. Manufacture of estolide from ricinoleic
acid with immobilized lipase. Japanese Kokai Tokkyo Koho 1993, JP 05211878
A 19930824.
[18] Yamaguchi C, Yoshida Y, Kawase M. Manufacture of estolide with immobilized
lipase. Japanese Kokai Tokkyo Koho 1993, JP 05304966 A 19931119.
[19] Yoshida Y, Kawase M, Yamaguchi C, Yamane T. Synthesis of estolides with
immobilized lipase. Yukagaku 1995;44:328–33.
[20] Goto M, Hatanaka C, Haraguchi T. Formation of ricinoleic acid by lipase immo-
bilized within isopropylacrylamide gel beads. Res Rep Kitakyushu Tech Coll
1999;32:161–4.
[21] Yoshida Y, Kawase M, Yamaguchi C, Yamane T. Enzymic synthesis of estolides
by a bioreactor. JAOCS 1997;74:261–7.
MMKOH molecular mass of potassium hydroxide [g mol−1
]
MS
OA
mass spectrometry
oleic acid, 9(Z)-octadecenoic acid
q(10S)-HOME (10S)-HOME productivity [g l−1 h−1
]
q(7S,10S)-DiHOME (7S,10S)-DiHOME productivity [g l−1 h−1
qOA
RA
RCF
t
]
OA consumption rate [g l−1 h−1
]
ricinoleic acid, 12(R)-hydroxy-9(Z)-octadecenoic acid
ratio competitive factor (1/˛)cis/(1/˛)trans [–]
time [h]
TOF
VKOH
WE
time-of-flight
volume of potassium hydroxide [ml]
wax esters
Wsample sample weight [g]
dry cell weight [g l−1
Y(10S)-HOME/X specific (10S)-HOME yield [g g−1
X
]
]
Y(7S,10S)-HOME/X specific (7S,10S)-DiHOME yield [g g−1
]
[22] Bódalo-Santoyo A, Bastida-Rodriguez J, Máximo-Martín MF, Montiel-Morte
MC, Murcia-Almagro MD. Enzymatic biosynthesis of ricinoleic acid estolides.
Biochem Eng J 2005;26:155–8.
[23] Bódalo A, Bastida J, Máximo MF, Montiel MC, Gómez M, Murcia MD. Production
of ricinoleic acid estolide with free and immobilized lipase from Candida rugosa.
Biochem Eng J 2008;39:450–6.
YP/X
ꢂ
˛
whole specific product yield [g g−1
increment [–]
competitive factor [–]
reaction yield [%]
]
ꢁ
[24] Klibanov AM. Improving enzymes by using them in organic solvents. Nature
2001;409:241–6.
[25] Horchani H, Bouaziz A, Gargouri Y, Sayari A. Immobilized Staphylococcus xylo-
sus lipase-catalysed synthesis of ricinoleic acid esters. J Mol Catal B: Enzym
2012;75:35–42.
[26] Hayes DG, Ye R. Optimization of the solvent-free lipase-catalyzed synthesis
of fructose-oleic acid ester through programming of water removal. JAOCS
2011;88:1351–9.
[27] Borgdorf R, Warwel S. Substrate selectivity of various lipases in the ester-
ification of cis and trans-9-octadecenoic acid. Appl Microbiol Biotechnol
1999;51:480–5.
[28] Stránsky´ K, Zarevúcka M, Valterová I, Wimmer Z. Gas chromatographic reten-
tion data of wax ester. J Chromatogr A 2006;1128:208–19.
[29] Tachi M, Iwamori M. Mass spectrometric characterization of cholesterol esters
and wax esters in epidermis of fetal, adult and keloidal human skin. Exp Der-
matol 2007;17:318–23.
[30] Rijpstra WIC, Reneerkens J, Piermas T, Damsté JS. Structural identification of
the b-hydroxy fatty acid-based diester preen gland waxes of shorebirds. J Nat
Prod 2007;70:1804–7.
[31] Vrkoslav V, Míková R, Cvacka J. Characterization of natural wax esters by
MALDI-TOF mass spectrometry. J Mass Spectrom 2008;44:101–10.
[32] Martinez E, Hamberg M, Busquets M, Díaz P, Manresa A, Oliw E. Biochemical
characterization of the oxygenation of unsaturated fatty acids by the diogenase
and hydroperoxide isomerase of Pseudomonas aeruginosa 42A2. J Biol Chem
2010;285:9339–45.
[33] Laane C, Boeren S, Vos K, Veeger C. Rules for optimization of biocatalysis in
organic solvents. Biotechnol Bioeng 1986;30:81–7.
[34] Aguieiras ECG, Veloso CO, Bevilaqua JV, Rosas DO, da Silva MAP, Langone
MAP. Estolides synthesis catalyzed by immobilized lipases. Enzyme Res
2011;2011:1–7.
Appendix B. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.
procbio.2012.12.006.
References
[1] Hayes DG, Kleiman R, Phillips BS. The triglyceride composition, structure,
and presence of estolides in the oils of Lesquerella and related species. JAOCS
1995;72:559–69.
[2] Burg DA, Kleiman R. Preparation of meadowfoam dimer acids and dimer esters,
and their use as lubricants. JAOCS 1991;68:600–3.
[3] Kleiman R, Spencer GF, Earle FR, Nieschlang HJ. Tetra-acid triglycerides con-
taining new hydroxy eicosadienoyl moiety in Lesquerella auriculata seed oil.
Lipids 1972;7:660–5.
[4] Smedley SR, Schroeder FC, Weibel DB, Meinwald J, Lafleur KA, Renwick JA, et al.
Labile defensive lipids from the glandular hairs of a caterpillar (Pieris rapae).
PNAS 2002;99:6822–7.
[5] Butovich IA. Cholesteryl esters as a depot for very long chain fatty acids in
human meibum. J Lipid Res 2009;50:501–13.
[6] Peláez M, Orellana C, Marqués A, Busquets M, Guerrero A, Manresa A. Natural
estolides produced by Pseudomonas sp. 42A2 grown on oleic acid: production
and characterization. JAOCS 2003;80:859–66.
[7] Cermak SC, Isbell TA. Synthesis and physical properties of mono-estolides with
varying chain lengths. Ind Crops Prod 2009;29:205–13.
[8] Achaya KT. Chemical derivatives of castor oil. JAOCS 1971;48:758–63.
[9] Cermak SC, Isbell TA. Synthesis of estolides from oleic and saturated fatty acids.
JAOCS 2001;78:557–65.
[35] Radzi SM, Basri M, Salleh AB, Ariff A, Mohammad R, Rahman MBA, et al. High
performance enzymatic synthesis of oleyl oleate using immobilised lipase from
Candida antarctica. Electron J Biotechnol 2005;8:291–8.
[36] Bayer E, Gfrörer P, Rentel C. Coordination-ionspray-MS (CIS-MS), a univer-
sal detection and characterization method for direct coupling with separation
techniques. Angew Chem Int Ed 1999;38:992–5.
[37] Nielen MWF. Maldi time-of-flight mass spectrometry of synthetic polymers.
Mass Spectrom Rev 1999;18:309–44.
[38] Kelly AR, Hayes DG. Lipase-catalyzed synthesis of polyhydric alcohol-
poly(ricinoleic acid) ester star polymers. J Appl Polym Sci 2006;101:1646–56.
[10] Isbell TA. Chemistry and physical properties of estolides. Grasas Aceites
2011;62:8–20.
[11] Cermak SC, Skender AL, Deppe AB, Isbell TA. Synthesis, physical properties of
tallow-oleic estolide 2-ethylhexyl esters. JAOCS 2007;84:449–56.
[12] Bódalo A, Bastida J, Máximo MF, Montiel MC, Murcia MD, Ortega S. Influ-
ence of the operating conditions on lipase-catalysed synthesis of ricinoleic acid
estolides in solvent-free systems. Biochem Eng J 2009;44: 214–9.