Hydroperoxide production from linoleic acid by heterologous Gaeumannomyces graminis tritici lipoxygenase: Optimization and scale-up

Chemical Engineering Journal 217 (2013) 82–90

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Hydroperoxide production from linoleic acid by heterologous Gaeumannomyces graminis tritici lipoxygenase: Optimization and scale-up Juan José Villaverde a,⇑, Vincent van der Vlist b, Sónia A.O. Santos a, Thomas Haarmann c, Kim Langfelder c, Minni Pirttimaa d, Antti Nyyssölä d, Sirpa Jylhä d, Tarja Tamminen d, Kristiina Kruus d, Leo de Graaff b, Carlos Pascoal Neto a, Mário M.Q. Simões e, M.R.M. Domingues e, Armando J.D. Silvestre a, Jasmin Eidner c, Johanna Buchert d a

CICECO and Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal Wageningen University, Laboratory of Systems and Synthetic Biology, Fungal Systems Biology, Dreijenplein 10, 6703 HB Wageningen, The Netherlands c AB Enzymes GmbH, Feldbergstrasse 78, D-64293 Darmstadt, Germany d VTT Biotechnology, P.O. Box 1500, FIN-02044 VTT, Finland e QOPNA and Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

" Hydroperoxides were produced from

"

" "

"

linoleic acid by Gaeumannomyces graminis lipoxygenase. Gaeumannomyces graminis lipoxygenase was obtained recombinantly in Trichoderma reesei. Yield and regioselectivity were optimized using 10 g/L linoleic acid. The process was investigated at industrially relevant substrate concentrations. At 300 g/L linoleic acid, the yield was 40% and the volumetric productivity 3.6 g/(L h).

a r t i c l e

i n f o

Article history: Received 20 August 2012 Received in revised form 15 November 2012 Accepted 20 November 2012 Available online 29 November 2012 Keywords: Lipoxygenase Linoleic acid Fatty acid hydroperoxides Optimization Scale-up

a b s t r a c t Linoleic acid was converted into hydroperoxides by a Gaeumannomyces graminis tritici lipoxygenase produced recombinantly in Trichoderma reesei. Hydroperoxide production was optimized using a facecentred experimental design in order to study the effects of pH, temperature and time on the conversion of linoleic acid into four regioisomeric hydroperoxyoctadecadienoic acids (HPODE): 13-(Z,E)-, 9-(E,Z)-, 13-(E,E)-, 9-(E,E)-HPODE. Fitting equations described satisfactorily the system behavior and showed that reaction time was the most influencing independent variable. A set of independent variables (pH = 6.7, temperature = 23.9 °C and time = 18 h) allowed to obtain high yields of hydroperoxides (88.0%) with a good selectivity for the 13-(Z,E)-HPODE isomer (47.4%) when the initial substrate concentration was 10 g/L. The production was further investigated using industrially relevant linoleic acid concentrations (100–300 g/L) leading to HPODE yields of 40% and the volumetric productivity 3.6 g/(L h), and a selectivity for 13-(Z,E)-HPODE of around 74%. Ó 2012 Elsevier B.V. All rights reserved.

⇑ Corresponding author. Present address: DTEVPF, Unit of Plant Protection Products, INIA, Ctra. de La Coruña, Km 7.5, 28040 Madrid, Spain. Tel.: +34 91 347 87 67; fax: +34 91 347 14 79. E-mail address: [email protected] (J.J. Villaverde). 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.11.090

1. Introduction Plant oils containing fatty acids and their glycerides are valuable and abundant renewable raw materials. In 2006, 127 million

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90

tons of plant oils were produced, of which approximately 79% were derived from palm, soybean, rapeseed and sunflower [1]. Many applications of vegetable oils and fatty acids, which include paints, coatings, adhesives, lubricants, detergents and plasticizers [2,3] take advantage of the their inherent physicochemical properties, but also of their chemical modification which takes place mainly in the ester/carboxylic moeity or in the unsaturated system. However, it has been claimed that transformations of fatty acids will be increasingly focusing on the alkyl chain instead of the carboxyl end [4,5]. Oxygenation of plant oils is an important route to specialty polymers and fine chemicals; and, in that perspective, unsaturated fatty acids from plant oils provide suitable targets for oxidative conversions [6]. In fact, fatty acid hydroperoxides could be interesting intermediates for the industrial valorization of plant oils. Fatty acids hydroperoxides are intermediates in the oxypolymerization of fatty acids involved in the cross-linking of drying oils in paint formulations [7] and versatile intermediates for the preparation of mid-chain hydroxy fatty acids [8], epoxides [9], alcohols [9], diacids [8] and oxoacids and aldehydes [8], which in turn can be used in polymer production. Chemical technologies to produce these types of intermediates from conventional lipid raw materials are not well developed and involve hazardous steps such as ozone oxidation [10] or environmentally harmful intermediates (e.g. bromoacids) [11]; or have low selectivity to obtain specific structural isomers and difficulty to stop the reaction without produce degradation products (e.g. Fenton’s reaction) [12,13]. Selective and efficient oxidation of unsaturated fatty acids to desired hydroperoxides can be achieved enzymatically using lipoxygenases as the catalysts. Lipoxygenases are iron or manganese containing enzymes, which catalyze the oxidation of polyunsaturated fatty acids containing cis,cis-1,4-pentadiene-type structures such as linoleic (C18:2), linolenic (C18:3) and arachidonic (20:4) acids into hydroperoxides (HPODEs). Linoleic acid is the most abundant lipoxygenase substrate in nature. It is present in 54% (by weight) in soybean oil and in 68% in sulflower oil [14]. ‘‘Tall oil fatty acid’’ (TOFA), which is derived from the pine oil by-product of kraft pulping process, is a non-crop source of linoleic acid. TOFA contains approximately 40–50% of linoleic acid [15]. Lipoxygenases have been found in plants [16], mammals [17] and microbes [18]. However, the most widely studied lipoxygenases are from soybean, having iron as the co-factor [19]. The only known manganese lipoxygenase is from the fungus Gaeumannomyces graminis (GGLOX) [18], which has been produced recombinantly using Pichia pastoris [20] and Trichoderma reesei [21] as the hosts. Enzymatic hydroperoxide production processes described so far use almost exclusively soybean lipoxygenases as the catalysts. Soybean lipoxygenases have been immobilized to various supports such as anion exchange resins [22], silica [23], and polyacrylamide [24]. Reactions have been carried out in aqueous phase [25] as well as in the presence of organic solvents in bi- or triphasic systems [26]. In addition, microemulsified reaction mixtures containing surfactants have been used [27]. To provide oxygen, the reactions have been carried out under oxygen pressure [28] or under stirring and oxygen sparging [25]. Despite the wide variety of production modes and conditions investigated, to our knowledge systematic multivariable analyses have not been used to study the enzymatic hydroperoxide production conditions. In addition, the substrate concentrations used have been low. In this work the conversion of linoleic acid to hydroperoxyoctadecadienoic acids (HPODEs) by GGLOX was investigated using an empirical face-centred second-order factorial design and the reaction conditions optimized. The empirical correlations of the process variables constructed allowed to predict the system behavior at lab scale, as a starting point for a subsequent scaling, and also to obtain the surface responses. Optimal conditions were

83

then used as the basis for experiments at industrially relevant substrate concentrations and for scaling up the process. 2. Materials and methods 2.1. Chemicals Linoleic acid (P99% purity), palmitic acid (BioXtra, P99%), ammonium iron(II) sulfate hexahydrate (ReagentPlusÒ, P98%), iron(II) chloride tetrahydrate (ReagentPlusÒ, 99%) and 2,6-ditert-butyl-4-methylphenol (BHT) (P99% purity) were supplied by Sigma–Aldrich. The scale-up experiments were performed using technical grade linoleic acid, which contained 64.7% linoleic acid. The remaining percentage was mainly oleic acid. Tween 20 (for molecular biology, Sigma–Aldrich) was used as an emulsifier in the lipoxygenase assays. Xylenol Orange disodium salt (grade p.a., for spectrophotometric determination of metal ions), hydrogen peroxide (30% purity) and formic acid (purity higher than 98%) were purchased from Fluka Chemie. Sulfuric acid (96% purity) was acquired from Acros Organics. HPLC-grade methanol, water, acetonitrile and chloroform were obtained from Fisher Scientific Chemicals. Solvents were filtered using a Solvent Filtration Apparatus 58061 (Supelco). 2.2. Lipoxygenase production The transformation of Trichoderma reesei with the Gaeumannomyces graminis tritici (GGLOX) lipoxygenase gene (AAK81882.1) has been described elsewhere [21]. The 30 L bioreactor (Sartorius Cplus) cultivation of the recombinant strain was carried out at 28.0 °C, at the initial pH 4.3, which later shifted to pH 5.3. The medium contained glucose, ammonium sulfate, monopotassium phosphate and a complex nitrogen source [29]. Culture supernatants were harvested and concentrated by ultrafiltration to give the final enzyme sample. The stability of GGLOX in an aqueous solution was tested, observing that is stable for at least 3 days, pH 8.0 and room temperature. In the presence of substrate (linoleic acid), the GGLOX is active at least after 24 h. 2.3. Analytical methods 2.3.1. Lipoxygenase activity assay The 25 mM linoleic acid solution containing 14 mg/mL Tween 20 was prepared as described previously [30] and used in the lipoxygenase assays. The reaction mixtures comprised 0.6 mM linoleic acid, McIlvaine buffer (pH 7.0) and the enzyme sample. Activity was determined by monitoring the change in absorbance at 234 nm. The extinction coefficient of 25,000 M1 cm1 for the linoleic acid hydroperoxides was used in the activity calculations [31]. Lipoxygenase activity is presented as mkat (mmol of HPODEs produced per second). 2.3.2. HPODE analyses Analyses of HPODEs, contained in a vial under N2 atmosphere, were carried out using a previously reported integrated procedure, by HPLC-UV-MS and the Ferrous Oxidation-Xylenol orange (FOX) method [12]. The equipment used was a Thermo Scientific Accela High Speed LC (Thermo Fisher Scientific, San Jose, CA, USA) composed of a variable loop Accela autosampler (200 vial capacity, set at a temperature of 15.0 °C), an Accela 600 LC pump and an Accela 80 Hz PDA detector. The optimal chromatographic conditions for the separation of linoleic acid oxidation products were achieved using a DiscoveryÒ C-18 (15 cm  2.1 mm  5 lm) column supplied by Supelco (Agilent Technologies, Waldbronn, Germany), at a flow rate of

84

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90

0.3 mL/min, with the column temperature set at 25.0 °C. The composition of the eluent was: eluent A: 0.1% of HCOOH in 90.8% of water and 9.1% of acetonitrile/eluent B: 0.1% of HCOOH in acetonitrile. Elution started with 70% of A and 30% of B for 5 min. After this the eluent composition was ramped to 70% of B in 25 min and finally to 90% of B in 2 min, which was maintained for 3 min. The column was re-equilibrated for 10 min before the next run. Prior to injection, each sample was dissolved in acetonitrile and filtered through a 0.2 lm PTFE syringe filter. The injection volume in the HPLC system was 10 lL and UV absorption at 234 nm was detected using the diode array detector. To confirm peak identification, the HPLC system was also coupled to a LCQ Fleet ion trap mass spectrometer (ThermoFinnigan, San Jose, CA, USA), equipped with an electrospray ionization source and operating in negative mode. The nitrogen sheath and auxiliary gas were 30 and 5 (arbitrary units), respectively. The spray voltage was 4.70 kV and the capillary temperature 300 °C, while capillary and tune lens voltages were set at 35 V and 125 V, respectively. The isolation width of precursor ions was 1.0 mass unit. The scan time was equal to 100 ms and the collision energy was optimized and set at 32 (arbitrary units), using helium as the collision gas. The detection was carried out considering a mass range of 100–2000 m/z and the CID-MS2 experiments were performed on mass-selected precursor ions in the range of m/z 85–320. The data acquisition was carried out using Xcalibur software (version 2.1.0 SP1 build 1160). The residual linoleic acid (%), contained in vials under N2 atmosphere, was determined after conversion into trimethylsilyl ester as described earlier [32,33] followed by gas-chromatographic determination (GC-FID) using the procedure described elsewhere [13]. Each analysis was carried out in triplicate and the presented results are the average of the obtained values (less than 5% variation). In the conversion experiments at high linoleic acid concentrations, the hydroperoxide concentrations were determined by UV spectroscopy at 234 nm from samples dissolved in ethanol (96%, v/v) as described above for the lipoxygenase activity assay. 2.3.3. Gel permeation chromatography (GPC) analysis of the reaction products The samples were dissolved in tetrahydrofuran to the approximate concentration of 10 mg/mL and analyzed using HP (Agilent) system 1050 HPLC equipped with a Waters Styragel HR4 E 8  300 mm column and an HP (Agilent) 1050 Autosampler. The detectors used were HP (Agilent) 1050 Diode Array UV Detector (250 nm) and HP Refractive Index analog detector. The UV detector was connected before the RI detector, which induced a slight shift (ca 30 s) in the elution times. Samples (injection volume 25 lL) were analyzed at ambient temperature. Tetrahydrofuran (inhibitor-free) was used as the mobile phase (0.5 mL/min). Polystyrenes were used as standards with molar masses of 114, 45, 12, 3.7 and 0.68 kDa. 2.4. Enzymatic conversion of linoleic acid into hydroperoxides 2.4.1. Optimization of reaction conditions in small scale The enzymatic oxidation reactions of linoleic acid were performed using Micro-Reaction Vessels (Supelco). The pH was adjusted using three different buffers to the following values: pH 4.0 (Na-acetate, 50 mM), pH 7.0 (Na-phosphate buffer, 50 mM) and pH 10.0 (borate-NaOH, 50 mM). For micellization, the linoleic acid was dissolved in CH2Cl2 in the reactor. The solvent was then evaporated under a stream of N2 to spread the linoleic acid onto the reactor walls, thus increasing its contact surface. The buffer was added to the reactor and the mixture was sonicated for 3 min. Then, the enzyme was added to the dispersion and with the experiments used for the factorial

design the final volume, enzyme dosage and LA concentration were fixed at 2.5 mL, 341 mkat/mL and 10 g/L, respectively. The mixture was left to react in the dark under magnetic stirring (900 rpm) and aeration through the liquid surface. The rest of variables were changed as follows: time, 2–18 h; temperature, 5.0–45.0 °C; pHvalue, 4.0–10.0. Additional experiments were carried out outside of the ranges of this factorial design using enzyme dosages, LA concentrations and reaction times between 3–1190 mkat/mL, 10– 320 g/L, 18–48 h, respectively. The reaction products were extracted with chloroform for GC and FOX analysis as follows. A 400 lL aliquot of the reaction mixture was collected under stirring (900 rpm), and 400 lL of chloroform was added. The sample was centrifuged (4100  g) for 3 min to achieve phase separation. 100 lL of the organic phase were introduced into a 1.5 mL amber glass vial together with 100 lL of chloroform with tetracosane (1.7 g/L). The sample was dried with N2 and stored at 20.0 °C under N2 atmosphere. The recovery factors were calculated for each pH, using palmitic acid as internal standard. The samples for the HPLC-UV-MS analysis were extracted the same way, but without adding tetracosane. 2.4.2. Reactions at high linoleic acid concentrations and scale-up of the process All experiments at high linoleic acid concentrations were carried out at 24.0 °C. The effect of linoleic acid concentration on the hydroperoxide yield was investigated in reaction mixtures containing 40 g/L, 100 g/L, 200 g/L and 300 g/L of linoleic acid, 24 mkat/mL of GGLOX and McIlvaine pH 7.0 buffer. Since the high concentrations of linoleic acid affected the pH, its solubility and foam formation, the pH of the reaction mixtures were adjusted to pH 7.6–7.8 with 10 M NaOH. The reaction mixtures were emulsified by sonication for 1 min per 5 mL before addition of the enzyme. The reaction mixtures were incubated in open 2 mL Eppendorf tubes, in the dark, at 24.0 °C and under shaking (1000 rpm., Eppendorf Thermomixer comfort). The impact of GGLOX concentration on hydroperoxide production at high linoleic acid concentration (300 g/L) was studied in a volume of 100 mL in baffled 250 mL Pyrex shake flasks at 25.0 °C. The emulsions were prepared by mixing 50 mL of 0.2 M McIlvaine pH 7.0 buffer with 30 g of linoleic acid and adding water to 98 mL. The mixture was micellized by sonication. GGLOX was added in dosages of 1–59 mkat/mL. The final volume was 100 mL and the final pH approximately 7.0. The effects of aeration and introduction of pure oxygen into the reaction mixture were compared at 100 g/L and at 300 g/L linoleic acid with 24 mkat/mL GGLOX as the catalyst. Borate-NaOH pH 9.5 buffer was added to the reaction mixture at the final concentration of 0.1 M. The final pH values of the emulsions were between pH 7.0–7.2. The emulsions were prepared by sonication as described above. The reaction was carried out in a graduated cylinder of 250 mL under magnetic stirring. Air and oxygen (99.5%) used were passed through water in order to avoid drying of the reaction mixtures. For efficient dispersion of the bubbles, air and oxygen streams were directed immediately above the magnet in the reaction mixture. A 1 L reaction was carried in order to study the distribution of HPODE isomers at a larger scale. The reaction mixtures contained 0.1 M McIlvaine pH 7.0 buffer (final pH 6.2), 300 g/L linoleic acid and 24 mkat/mL of GGLOX. The components were mixed in a glass vessel without sonication and the mixture was continuously stirred using a Teflon mixer at 400 rpm. 2.5. Experimental design and statistical analysis A face-centred second order factorial experimental design [34,35] was chosen to examine and quantify the effect of the

85

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90

DV ¼ b0 þ

3 3 X X X bi X i þ bii X 2i þ bij X i X j i¼1

i
i¼1

where DV represents each of the dependent variables (system responses) studied, Xi, Xj are the normalized variables previously defined, and b0, bi and bij are the fitting parameters (obtained by multiple regression between the system responses (DV) and the values of the normalized variables) by which, due to normalization, it was possible to determine and compare the effects of each of the independent variables on the dependent variables. Regressions and the statistical analysis were performed using the Excel statistical module.

9-(E,E)-HPODE

13-(E,E)-HPODE

3.1. Effect of enzyme dosage on HPODE generation at low substrate concentration The effect of enzyme dosage (3–511 mkat/mL) on HPODE generation at 10 g/L linoleic acid was investigated by a series of preliminary experiments at 25.0 °C and at pH 7.0 (Fig. 1). As expected, without enzyme addition, autoxidation of linoleic acid was observed and 12.1% of the initial linoleic acid was converted into HPODEs in the reference sample after 18 h, while after 48 h the HPODE yield was as high as 38.7%. When GGLOX was used as catalysts, the HPODE yields were around 70–85% after 18 h for doses between 170 and 511 mkat/mL (Fig. 1), being the highest catalyzed effect 71.8%, which was achieved using 341 mkat/mL GGLOX. The results obtained show that with longer incubation times and increasing enzyme doses, lower yields of HPODEs were obtained. HPODEs are produced more rapidly at higher enzyme 90

2h

18 h

48 h

80 70

HPODE%

60 50 40 30 20

13-(Z,E)-HPODE

90

80

70

60

50

40 2

18 Reaction time (hours)

3. Results and discussion

9-(E,Z)-HPODE

100

Relative HPODE isomers %

biological process variables on the parameters defining the oxidation progress (linoleic acid%, 13-(Z,E)-HPODE%, 9-(E,Z)-HPODE%, 13-(E,E)-HPODE% and 9-(E,E)-HPODE%) (see Section 2.4.1.). To compare the effects of the independent variables (pH, 4.0–10.0; Temperature, 5.0–45.0 °C and time, 2–18 h) their values were normalized. Experimental data were fitted to polynomial models of the type:

48

3

511

Enzyme dosage (mkat/mL)

Fig. 2. Relative proportion of HPODE isomers (%) generated by GGLOX at 34 mkat/ mL during 2, 18 and 48 h (left), together with those generated during 2 h at 3 and 511 mkat/mL (right).

concentrations, and therefore the higher HPODE concentration in early reaction stages give them more time to degrade during the treatment into low molecular weight compounds [36] and/or oligomeric structures [13]. The selectivity of the lipoxygenase reaction towards the formation of different positional isomers was determined by HPLC [12]. After a 2 h incubation period and 34 mkat/mL of GGLOX, 13(Z,E)-HPODE represented 70% of the total HPODEs formed, whereas 9-(E,Z)-HPODE, 13-(E,E)-HPODE and 9-(E,E)-HPODE accounted for 14%, 9% and 7%, respectively. This trend towards the formation of 13-(Z,E)-HPODE was also observed at different doses of GGLOX. Finally, it has been reported that 13-HPODEs are also the predominant products when native GGLOX was used [18]. Furthermore, with increasing incubation times, selectivity towards 13-HPODE generation clearly decreased (e.g. at the dose of 34 mkat/mL, Fig. 2). This is in agreement with previously published studies [37,38] and is most likely a result of autoxidation. Indeed, autoxidation was clearly less selective and resulted in the formation of 33% of 13-(Z,E)-HPODE, 25% of 9-(E,Z)-HPODE, 22% of 13-(E,E)HPODE and 20% of 9-(E,E)-HPODE. Finally, with increasing enzyme doses (e.g. at doses of 3 and of 511 mkat/mL, Fig. 2) a higher production of the isomer 13-(Z,E)HPODE is observed, suggesting that the enzymatic reaction is predominating over autoxidation. It can be concluded that the reaction conditions greatly influence the conversion of linoleic acid into HPODEs and the selectivity in the GGLOX catalyzed reactions. Furthermore, synergic effects can also affect the reaction. For example pH and temperature affect both the solubility of the substrate and the activity and stability of the enzyme. Therefore, the study on the effects of various combinations of experimental conditions was carried out, by factorial design of experiments.

10

3.2. Optimization of experimental conditions by factorial design

0 0

3

17

34

85

170

341

511

Enzyme dosage (mkat/mL) Fig. 1. HPODE formation (% of initial linoleic acid) at different doses of GGLOX (3– 511 mkat/mL), at 10 g/L linoleic acid, at 25.0 °C and at pH 7.0.

After these preliminary experiments, factorial design was used to analyze the effects of the three independent variables: pH, temperature, and time on the dependent variables or system responses, linoleic acid%, 13-(Z,E)-HPODE%, 9-(E,Z)-HPODE%,

86

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90

Table 1 Structure of the experimental design and results of linoleic acid oxidation by GGLOX (Exp. 1–18), together with deviations estimated by the polynomials models (in brackets), and additional experiments to validate the models (Exp. 19 and 20). The linoleic acid (LA) concentration and GGLOX dose were fixed at 10 g/L and 341 mkat/mL, respectively. Exp.

pH

Temp. (°C)

Time (h)

LA%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

4.0 4.0 4.0 4.0 10.0 10.0 10.0 10.0 7.0 7.0 7.0 7.0 4.0 10.0 7.0 7.0 7.0 7.0 8.5 5.5

5.0 5.0 45.0 45.0 5.0 5.0 45.0 45.0 25.0 25.0 25.0 25.0 25.0 25.0 5.0 45.0 25.0 25.0 35.0 15.0

2 18 2 18 2 18 2 18 10 10 10 10 10 10 10 10 2 18 14 6

43.2 40.3 70.3 24.9 66.6 20.9 86.5 80.8 27.9 25.7 28.2 27.5 52.3 66.5 32.3 52.6 45.1 12.9 40.3 32.2

(2.5) (1.1) (2.0) (0.5) (1.6) (0.1) (1.0) (0.4) (3.6) (5.8) (3.3) (4.0) (6.1) (2.3) (5.2) (3.2) (7.1) (1.3) (3.9) (2.4)

13-(Z,E)-HPODE%

9-(E,Z)-HPODE%

13-(E,E)-HPODE%

9-(E,E)-HPODE%

Total-HPODE%

9.4 19.1 14.6 28.6 12.3 33.0 7.3 7.1 38.9 42.3 40.2 39.3 27.1 24.1 23.6 19.7 28.8 48.7 31.7 28.6

1.3 4.0 3.5 16.3 2.7 7.9 2.0 2.9 10.6 12.2 10.4 10.7 7.1 3.1 5.8 8.0 4.2 11.7 9.8 7.5

0.5 6.6 5.3 17.8 1.5 11.9 1.5 3.5 9.2 7.2 7.6 7.8 6.2 3.1 7.1 9.8 9.9 11.7 7.2 5.5

0.5 7.0 5.8 14.7 1.8 13.2 1.6 3.8 12.0 10.0 11.1 10.7 7.1 3.7 8.0 9.4 9.4 11.7 8.9 8.1

11.6 36.7 29.2 77.4 18.3 65.9 12.4 17.3 70.8 71.7 69.3 68.5 47.6 34.0 44.6 46.8 52.3 83.9 57.6 49.7

(1.4) (0.4) (1.4) (0.4) (1.4) (0.4) (1.4) (0.4) (0.8) (4.2) (2.1) (1.2) (2.1) (2.1) (2.1) (2.0) (5.6) (1.5) (0.1) (0.9)

13-(E,E)-HPODE% and 9-(E,E)-HPODE%, in order to identify the key variables and possible synergic effects, aiming to optimize conversion into HPODE and to maximize the selectivity for 13-(Z,E)HPODE. Table 1 shows the structure of the experimental design, together with the results obtained in the experiments for the dependent variables. For experiments 1, 2, 5, 6 and 15 the global mass balance was not closed, most probably due to precipitation of linoleic acid at low temperature (5.0 °C). In fact, part of the substrate remained visibly stuck to the vessel wall leading to variation in the amount of linoleic acid accessible to the enzyme. Experimental results were fitted by means of least square multiple regression (Table 2). The models obtained allowed to predict the experimental results with errors lower than 7.1% for residual linoleic acid and 5.6%, 2.6%, 2.7% and 2.6% for the generation of 13-(Z,E)-HPODE%, 9-(E,Z)-HPODE%, 13-(E,E)-HPODE% and 9-(E,E)HPODE%, respectively (Table 1). The models were validated conducting two additional experiments (Table 1, entries 19 and 20). The deviations in the prediction of the results were quite small (Table 1), confirming the accuracy of the models. Table 3 shows the optimum values of the independent variables, predicted by the empirical correlations of the process variables, to obtain the highest HPODE yields: Row 1 shows the best conditions to obtain the highest HPODE% together with the highest selectivity to the 13-(Z,E)-HPODE isomer and Row 2 shows the best conditions for the maximum HPODE yield. There are very little differences between the selectivities for 13-(Z,E)-HPODE formation (0.7%) and between the HPODE yields obtained (0.3%) under these conditions. The temperature variable is significant at the 95% confidence level in the decrease of 13-(Z,E)-HPODE production (Table 2). Above 27.0 °C, the production of 13-(Z,E)-HPODE decreases (Fig. 3) and the conversion into 9-(E,Z)-HPODE and 13-(E,E)-HPODE increases moderately (drop in GGLOX selectivity) with a concomitant decrease in the total HPODE yield. The increase of pH and temperature at 18 h reaction time (Fig. 3) results into a decrease of the yield of the 13-(Z,E)-HPODE isomer. The maximum observed in this response surface come from the two significant quadratic terms, b11 and b22, presented in the equation for the 13-(Z,E)-HPODE% (Table 2). These quadratic terms are also present in the equation for the total HPODE%.

(0.4) (0.0) (0.4) (0.0) (0.7) (0.3) (0.7) (0.3) (1.0) (2.6) (0.8) (1.1) (0.7) (2.0) (1.3) (1.3) (2.2) (0.5) (1.0) (0.9)

(0.6) (0.6) (0.8) (0.4) (0.6) (0.6) (0.8) (0.4) (0.9) (1.1) (0.7) (0.5) (0.4) (0.3) (0.0) (0.7) (2.7) (2.0) (0.7) (0.1)

(0.4) (0.6) (0.5) (0.4) (0.1) (0.9) (0.2) (0.7) (1.7) (0.3) (0.8) (0.4) (0.0) (1.2) (0.8) (0.4) (1.4) (2.6) (0.4) (0.9)

(0.7) (0.7) (0.5) (0.4) (1.4) (1.4) (1.1) (1.1) (4.4) (5.3) (2.9) (2.1) (2.3) (5.0) (4.1) (3.1) (3.7) (3.6) (0.0) (0.9)

Reaction time is the most significant independent variable in all polynomial models. Within the range studied, increasing reaction time results in a higher yield of HPODEs. This suggests that active enzyme was still present at the end of the reactions. 3.3. Reactions at high substrate concentrations An industrially feasible process requires a substantially higher substrate concentration than that used in the above described optimization studies (Table 3). The first experiments at substrate concentrations between 40 and 300 g/L of linoleic acid (Table 4) were carried out in the optimal conditions determined at lab scale for 10 g/L of linoleic acid (Row 1, Table 3). However the low solubility of the linoleic acid at these higher concentrations, leading to foam formation, affected drastically the reaction conditions. To avoid foam formation the pH had to be raised still within the ranges where de enzyme was active (pH 7.6–7.8), with enzyme dosage of 24 mkat/mL for 7 h at 24.0 °C (Table 4). The results indicate that the hydroperoxide concentrations achieved at different linoleic acid concentrations were of the same order, with the highest value being obtained at 100 g/L linoleic acid (86 mM HPODE). However, the conversion yields were considerably lower than those reported above with a maximum of 34% for 40 g/L of linoleic acid and decreasing drastically above 100 g/L. These results could be a consequence both from the lack of efficient mixing at such high concentrations, which was previously demonstrated [25] but also from the lack of oxygen, which in the reaction mixture is obviously an important factor affecting lipoxygenation rate. The effects of aeration and introduction of pure oxygen into the reaction mixture were compared at 100 g/L and at 300 g/L linoleic acid with 24 mkat/mL GGLOX. Under air sparging and linoleic acid concentration of 100 g/L (10-fold higher than the optimization studies above), the HPODE yields were 34% and 50% after 7 h and 24 h, respectively. When the substrate concentration was increased to 300 g/L the resulting HPODE yields were 46% and 58% after 7 h and 24 h, respectively. Finally, there was no clear difference between the HPODEs yields under oxygen or under air atmosphere, although under aeration the reaction was initiated slightly slower than under oxygen sparging, as a lag phase could be detected in the former case (Fig. 4A). So, and in conclusion, good yields (58%) can be obtained with aeration at 300 g/L after 24 h of reaction.

87

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90 Table 2 Regression parameters for each variable, goodness of fit and signification of regression equations. Fitting parameters b0 b1 b2 b3 b12 b13 b23 b123 b11 b22 b33 R2 a b

Name

pH Temp time

LA% coefficient

13-(Z,E)-HPODE% coefficient

9-(E,Z)-HPODE% coefficient

13-(E,E)-HPODE% coefficient

9-(E,E)-HPODE% coefficient

Total-HPODE% coefficient

31.52a 9.03a 11.18a 13.19a 8.51a 0.39 0.31 10.31a 23.69a 6.74b 6.71b 0.9728

38.12a 1.50 2.01 6.41a 5.70a 0.40 2.08 3.15a 10.47a 14.42a 2.68 0.9704

9.63a 1.36b 1.10 2.91a 2.53a 1.18 0.73 1.80a 3.18a 1.38 0.33 0.9239

8.32a 1.49a 1.03b 3.28a 3.05a 0.78 0.25 1.85a 4.03a 0.23 2.12b 0.9431

10.32a 1.10b 0.48 3.13a 2.83a 0.23 0.85 1.45a 4.29a 0.99 0.86 0.9440

66.43a 5.46a 0.60 15.74a 14.10a 2.60 2.45 8.23a 21.98a 17.08a 5.32 0.9842

Indicates a significance level of <0.05. Indicates a significance level of <0.10.

The effect of different GGLOX dosages on the hydroperoxide formation was investigated at a substrate concentration of 300 g/L, in baffled shake flasks. The reactions proceeded at a linear rate for 24 h (data not shown). The yields obtained at the end of 7 h and 24 h are presented in Fig. 4B. The results indicate that the yield increased when increasing enzyme concentration until 24 mkat/mL,

above which the yield (approximately 50%) reached a plateau. Presumably the dissolution rate of linoleic acid into the aqueous phase was limiting the reaction at this point, instead of the true catalytic rate of the enzyme. It has been suggested that the hydroperoxides formed gather on the surfaces of the oil droplets, which may limit the access of the enzyme to the substrate [25].

Table 3 Predicted experimental conditions and results obtained using the empirical correlations of the process variables. Row 1 shows the conditions to obtain the highest selectivity for 13-(Z,E)-HPODE. Row 2 shows the conditions to obtain the maximum conversion into HPODEs. Predicted conditions and yields Row

1 2

Experimental conditions

Yields

pH

Temp. (°C)

Time (h)

LA%

13-(Z,E)-HPODE%

9-(E,Z)-HPODE%

13-(E,E)-HPODE%

9-(E,E)-HPODE%

Total-HPODE%

6.7 6.3

23.9 27.0

18 18

10.5 11.6

47.4 46.7

12.3 12.9

13.8 14.2

14.4 14.4

88.0 88.3

Fig. 3. Influence of temperature and pH on the 13-(Z,E)-HPODE yield. Reaction time, 18 h; linoleic acid concentration, 10 g/L and GGLOX dose, 341 mkat/mL.

88

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90

Table 4 Effect of linoleic acid concentration on the HPODE yield.a Linoleic acid concn. (g/L)

HPODE concn. (mM)

HPODE yield (% (mol/mol))

40 100 200 300

49 86 64 57

34 24 9 5.3

a Experimental conditions: pH 7.6–7.8.; T = 24.0 °C; GGLOX = 24 mkat/mL GGLOX. The errors were within 15%.

t = 7 h;

dosage

Fig. 5. Hydroperoxide production in large scale in a 1 L reaction mixture containing 300 g/L linoleic acid, 0.1 M McIlvaine buffer pH 7.0 and 0.3 g/L GGLOX. The emulsion pH was 6.2 after addition of GGLOX. The pH of the reaction mixtures were adjusted to 7.6–7.8 with 10 M NaOH. Total HPODE yield determined by the FOX assay (closed circles) and by UV-spectroscopy (open circles). Isomer yields: 13(Z,E)-HPODE (closed squares), 9-(E,Z)-HPODE (closed triangles), 13-(E,E)-HPODE (open diamonds) and 9-(E,E)-HPODE (crosses).

Fig. 6. GPC elution curves of the products formed in the GGLOX catalyzed reactions in 1 L scale using UV (250 nm). For molar mass scale reference, a series of PPS standards is shown in the UV detection case. Fig. 4. (A) Hydroperoxide formation from linoleic acid at 100 g/L (squares) and 300 g/L (circles) of linoleic acid by GGLOX under oxygen (filled symbols) or air (open symbols) sparging. The final pH values of the emulsions were between pH 7.0–7.2 and the enzyme dosage used was 24 mkat/mL. (B) Effect of enzyme dosage on hydroperoxide formation in reaction mixtures containing 300 g/L linoleic acid and McIlvaine buffer 0.1 M, pH 7.0. Yields at 7 h (triangles) and at 24 h (squares).

3.4. Scale up of the enzymatic reaction The GGLOX catalyzed HPODE production was further scaled up to a 1 L reactor. The final HPODE yield after 27 h was 280 mM corresponding to a 40% conversion (measured both by UV and FOX) of the linoleic acid and to a volumetric productivity of 3.6 g/(L h) (Fig. 5). The individual hydroperoxide isomers were further analyzed by HPLC–UV–MS. After 24 h HPODEs comprise 29.9% of 13(Z,E)-HPODE, 5.9% of 9-(E,Z)-HPODE, 2.3% of 13-(E,E)-HPODE and

2.5% of 9-(E,E)-HPODE (Fig. 5). Hence, 13-(Z,E)-HPODE represented 74% of the HPODE products. A limitation of the UV-spectroscopy analysis of the total HPODEs is that both monomers and oligomers with conjugated double bonds absorb [13,39]. However, the UV analysis and the FOX analysis gave very similar results, suggesting that no polymerization of HPODEs to compounds with conjugated double bonds had taken place. The HPODE yields obtained from this 1 L reaction (Fig. 5) were found to be coherent with the predictions obtained using the empirical correlations of the process variables (Table 3). The molar masses of the products of the GGLOX catalyzed reactions were further analyzed by GPC using UV detector. In addition to define the molar mass region of the samples, the increase of UV response as a function of reaction time could also be used for verifying the formation of conjugated double bond systems as a consequence of HPODE formation.

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90

The results from the GPC analysis are presented in Fig. 6. Slight broadening of the elution peak was visible after long reaction time, suggesting that minor polymerization may have taken place. The peaks eluting after 20 min are system peaks. The GPC results are in accordance with the comparison between the results obtained using the FOX assay and the results from the UV-spectroscopic assay of the total HPODEs. Both analyses suggested that extensive polymerization did not take place under the conditions used. Reports on the enzymatic conversion of linoleic acid into hydroperoxides at industrially relevant concentrations (or even in preparative scale) are scarce in the literature. In a process carried out under 10 bar oxygen pressure, 56 g/L of linoleic acid was subjected to lipoxygenase-1 catalyzed oxidation. A yield of 40% was achieved in this process with 85% of the hydroperoxide isomers being 13-HPODEs [38]. In another report a biphasic system with octane as solvent was used for converting 40 g/L of linoleic acid into 13-HPODE (representing 92% of the hydroperoxides) under oxygen sparging in the presence of lipoxygenase-1. A volumetric productivity of 12 g/(L h) and a yield of 30% was obtained in the process [26]. To our knowledge the only attempt to use lipoxygenase substrate concentrations over 100 g/L was reported for the oxidation of flax-seed oil hydrolysate containing 55% linolenic acid and 14% linoleic acid. In this process 143 g/L of flax-seed oil was oxidized by fractionated soybean seed extract. Oxygen saturation was achieved by agitation. The linolenic and linoleic acids fractions were converted into hydroperoxides with 64% and 45% yields, respectively. The total amount of fatty acids transformed corresponded to a volumetric productivity of 60 g/(L h). In this process the regioselectivity was also high, with more than 80% of 13HPODE [25]. In the present work, a substrate, linoleic acid, at a considerably higher concentration was used (300 g/L). The yield (40%) and selectivity (74% for 13-(Z,E)-HPODE) obtained were comparable to the previously reported values for the flax-seed oil hydrolysate, although with a slightly lower volumetric productivity of 3.6 g/ (L h). The present process could most likely be improved by increasing the accessibility of the substrate by detergent addition or by modification of the geometry and mixing of the vessel.

4. Conclusions Hydroperoxide production from linoleic acid by a Gaeumannomyces graminis tritici lipoxygenase was optimized, using a facecentred experimental design, in order to study the effects of pH, temperature and time on the conversion into four regioisomeric hydroperoxides: 13-(Z,E)-, 9-(E,Z)-, 13-(E,E)-, 9-(E,E)-HPODE. The empirical correlations of the process variables for the GGLOX system were found to fit well with the experimental results, and the optimal experimental conditions predicted were successfully confirmed by several assays. At low substrate concentrations, the GGLOX treatment results in fatty acid hydroperoxides yields as high as 85%. The main isomer produced was always 13-(Z,E)HPODE being reaction time the most influencing independent variable. A set of independent variables (pH = 6.7, temperature = 23.9 °C and time = 18 h) allowed to obtain high yields of hydroperoxides (88.0%) with a good selectivity for the 13-(Z,E)HPODE isomer (47.4%) when the initial substrate concentration was 10 g/L. The production was further investigated using industrially relevant linoleic acid concentrations. At 300 g/L of linoleic acid the hydroperoxides yield was approximately 40% and the volumetric productivity 3.6 g/(L h). Selectivity was good with 74% of the products being 13-(Z,E)-HPODE. Sparging with pure oxygen is not needed. The use of normal air for aeration is enough which lowers the costs of the process.

89

Acknowledgments The authors gratefully acknowledge the financial support provided by the European Research Project (Novel enzyme tools for production of functional oleochemicals from unsaturated lipids (ERA-NOEL), ERA-IB/BIO/0001/2008), CICECO (Pest-C/CTM/ LA0011/2011) and QOPNA (PEst-C/QUI/UI0062/2011). Juan José Villaverde and Sonia A.O. Santos also thank FCT-Fundação para a Ciência e a Tecnologia for the awarding of a postdoctoral grant (BPD/UI89/4520/2009) and a Ph.D. Grant (SFRH/BD/42021/2007), respectively. Martti Alkio is thanked for the GPC analyses.

References [1] Anonymous, ProdSTAT: Crops, Food and Agriculture Organization of the United Nations, Rome, 2007. [2] M.N. Belgacem, A. Gandini, Materials from vegetable oils: major sources, properties and applications, in: M.N. Belgacem, A. Gandini (Eds.), Monomers, Polymers and Composites from Renewable Resources, Elsevier, 2008, pp. 39– 66. [3] H.A. Hamid, R. Yunus, U. Rashid, T.S.Y. Choong, A.H. Al-Muhtaseb, Synthesis of palm oil-based trimethylolpropane ester as potential biolubricant: chemical kinetics modeling, Chem. Eng. J. 200–202 (2012) 532–540. [4] J.O. Metzger, U. Bornscheuer, Lipids as renewable resources: current state of chemical and biotechnological conversion and diversification, Appl. Biochem. Biotechnol. 71 (2006) 13–22. [5] E. Santacesaria, R. Tesser, M. Di Serio, R. Turco, V. Russo, D. Verde, A biphasic model describing soybean oil epoxidation with H2O2 in a fed-batch reactor, Chem. Eng. J. 173 (2011) 198–209. [6] A. Köckritz, A. Martin, Oxidation of unsaturated fatty acid derivatives and vegetable oils, Eur. J. Lipid Sci. Technol. 110 (2008) 812–824. [7] R. Gorkum, E. Bouwman, The oxidative drying of alkyd paint catalysed by metal complexes, Coord. Chem. Rev. 249 (2005) 1709–1728. [8] H.W. Gardner, Oxygen radical chemistry of polyunsaturated fatty acids, Free Radic. Biol. Med. 7 (1989) 65–86. [9] E.N. Frankel, Lipid Oxidation, second ed., The Oily Press, Bridgwater, England, 2005. [10] A. Leitzke, C. Sonntag, Ozonolysis of unsaturated acids in aqueous solution: acrylic, methacrylic, maleic, fumaric and muconic acids, Ozone Sci. Eng. 31 (2009) 301–308. [11] S.D. Richardson, M.J. Plewa, E.D. Wagner, R. Schoeny, D.M. DeMarini, Occurrence, genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in drinking water: a review and roadmap for research, Mutat. Res. 8 (1993) 127–131. [12] J.J. Villaverde, S.A.O. Santos, M.M.Q. Simões, C. Pascoal Neto, M.R.M. Domingues, A.J.D. Silvestre, Analysis of linoleic acid hydroperoxides generated by biomimetic and enzymatic systems through an integrated methodology, Ind. Crops Prod. 34 (2011) 1474–1481. [13] J.J. Villaverde, S.A.O. Santos, E. Maciel, M.M.Q. Simões, C. Pascoal Neto, M.R.M. Domingues, A.J.D. Silvestre, Formation of oligomeric alkenylperoxides during the oxidation of unsaturated fatty acids: an electrospray ionization tandem mass spectrometry study, J. Mass Spectrom. 47 (2012) 163–172. [14] F.D. Gunstone, J. Alander, S.Z. Erhan, B.K. Sharma, T.A. McKeon, J.-T. Lin, Nonfood uses of oils and fats, in: F.D. Gunstone, J.L. Harwood, A.J. Dijkstra (Eds.), The Lipid Handbook, third ed., Taylor & Francis Group, Boca Raton, FL, 2006, pp. 591–636. [15] S. Preis, R.G. Herman, J.S. Lavelle, J.E. Roberts, Radiation curable coating formulations comprising polyesters. US2007043135, 2007. [16] J.N. Siedow, Plant lipoxygenase: structure and function, Annu. Rev. Plant Physiol. Plant Mol. Biol. 42 (1991) 145–188. [17] P. Borgeat, M. Hamberg, B. Samuelsson, Transformation of arachidonic acid and homo-gamma-linolenic acid by rabbit polymorphonuclear leukocytes. Monohydroxy acids from novel lipoxygenases, J. Biol. Chem. 251 (1976) 7816– 7820. [18] C. Su, E.H. Oliw, Manganese lipoxygenase. Purification and characterization, J. Biol. Chem. 273 (1998) 13072–13079. [19] A.R. Brash, Lipoxygenases: occurrence, functions, catalysis, and acquisition of substrate, J. Biol. Chem. 274 (1999) 23679–23682. [20] M. Cristea, Å. Engström, C. Su, L. Hörsten, E.H. Oliw, Expression of manganese lipoxygenase in Pichia pastoris and site-directed mutagenesis of putative metal ligands, Arch. Biochem. Biophys. 434 (2005) 201–211. [21] A. Nyyssölä, R. Heshof, T. Haarmann, J. Eidner, A. Westerholm-Parvinen, K. Langfelder, K. Kruus, L. de Graaff, J. Buchert, Methods for identifying lipoxygenase producing microorganisms on agar plates, AMB Express 2 (2012) 17. [22] S. Kermasha, N. Dioum, B. Bisakowski, M. Vega, Biocatalysis by immobilized lipoxygenase in a ternary micellar system, J. Mol. Catal. B: Enzyme 19–20 (2002) 305–317. [23] A. Karout, C. Chopard, A.C. Pierre, Immobilization of a lipoxygenase in silica gels for application in aqueous media, J. Mol. Catal. B: Enzyme 44 (2007) 117– 127.

90

J.J. Villaverde et al. / Chemical Engineering Journal 217 (2013) 82–90

[24] M.C. Pinto, P. Macías, Synthesis of linoleic acid hydroperoxide using immobilized lipoxygenase in polyacrylamide gel, Appl. Biochem. Biotechnol. 59 (1996) 309–318. [25] M.L. Fauconnier, M. Marlier, An efficient procedure for the production of fatty acid hydroperoxides from hydrolyzed flax seed oil and soybean lipoxygenase, Biotechnol. Tech. 10 (1996) 839–844. [26] P. Drouet, D. Thomas, M.D. Legoy, Production of 13(S)-hydroperoxy-9(Z),11(E)octadecadienoic acid using soybean lipoxygenase 1 in a biphasic octane–water system, Tetrahedron Lett. 35 (1994) 3923–3926. [27] G.J. Piazza, Lipoxygenase catalyzed hydroperoxide formation in microemulsions containing nonionic surfactant, Biotechnol. Lett. 14 (1992) 1153–1158. [28] G. Iacazio, G. Langrand, J. Baratti, G. Buono, C. Triantaphylidés, Preparative, enzymatic synthesis of linoleic acid (13S)-hydroperoxide using soybean lipoxygenase-1, J. Org. Chem. 55 (1990) 1690–1691. [29] T. Nakari-Setälä, M. Paloheimo, J. Kallio, J. Vehmaanperä, M. Penttilä, M. Saloheimo, Genetic modification of carbon catabolite repression in Trichoderma reesei for improved protein production, Appl. Environ. Microbiol. 75 (2009) 4853–4860. [30] G.E. Anthon, D.M. Barrett, Colorimetric method for the determination of lipoxygenase activity, J. Agric. Food Chem. 49 (2001) 32–37. [31] M.R. Egmond, M. Brunori, P.M. Fasella, The steady-state kinetics of the oxygenation of linoleic acid catalysed by soybean lipoxygenase, Eur. J. Biochem. 61 (1976) 93–100.

[32] C.S.R. Freire, A.J.D. Silvestre, C.P. Neto, Identification of new hydroxy fatty acids and ferulic acid esters in the wood of Eucalyptus globulus, Holzforschung 56 (2002) 143–149. [33] J.J. Villaverde, R.M.A. Domingues, C.S.R. Freire, A.J.D. Silvestre, C. Pascoal Neto, P. Ligero, A. Vega, Miscanthus x giganteus extractives: a source of valuable phenolic compounds and sterols, J. Agric. Food Chem. 57 (2009) 3626–3631. [34] G.E.P. Box, N.R. Draper, Empirical model-building and response surfaces, John Wiley & Sons, New York, 1987. [35] L. Mesa, E. González, C. Cara, M. González, E. Castro, S.I. Mussatto, The effect of organosolv pretreatment variables on enzymatic hydrolysis of sugarcane bagasse, Chem. Eng. J. 168 (2011) 1157–1162. [36] H. Esterbauer, Aldehydic products of lipid peroxidation, in: D.C.H. McBrien, T.F. Slater (Eds.), Free Radicals, Lipid Peroxidation and Cancer, Academic Press, London, 1982, pp. 101–128. [37] C. Dufour, M. Loonis, Regio- and stereoselective oxidation of linoleic acid bound to serum albumin: identification by ESI–mass spectrometry and NMR of the oxidation products, Chem. Phys. Lipids 138 (2005) 60–68. [38] D. Martini, G. Iacazio, D. Ferrand, G. Buono, C. Triantaphylidés, Optimization of large scale preparation of 13-(S)-hydroperoxy-9Z,11E-octadecadienoic acid using soybean lipoxygenase. Application to the chemoenzymatic synthesis of (+)-coriolic acid, Biocatalysis 11 (1994) 47–63. [39] K. Miyashita, K. Fujimoto, T. Kaneda, Formation of dimers during the initial stage of autoxidation in methyl linoleate, Agric. Biol. Chem. 46 (1982) 751– 755.