Journal of Molecular Catalysis B: Enzymatic 72 (2011) 1319

Contents lists available at ScienceDirect

Journal of Molecular Catalysis B: Enzymatic

Application of response surface design to solvent, temperature and lipase selection for optimal monoglyceride production
Diana Marcela Cetina a , Gloria Ins Giraldo b , Carlos Eduardo Orrego c,
a Instituto de Biotecnologa y Agroindustria, Departamento de Ingeniera Qumica, Universidad Nacional de Colombia Sede Manizales, Campus la Nubia Km 4 Va al Magdalena, AA 127, Manizales, Colombia b Departamento de Fsica y Qumica, Universidad Nacional de Colombia Sede Manizales, Cra. 27 No. 64-60, Manizales, Colombia c Instituto de Biotecnologa y Agroindustria, Departamento de Fsica y Qumica, Universidad Nacional de Colombia Sede Manizales, Campus la Nubia Km 4 Va al Magdalena, AA 127, Manizales, Colombia

a r t i c l e

i n f o

a b s t r a c t
This work focuses on the use of a temperature and solvent lipase stability procedure as a practical approach for selection of the most favorable conditions for the enzyme catalysis production of monoglycerides from glycerin and triolein. Two lipases were selected for analysis: a lipase from Candida rugosa immobilized on chitosan and a lipase from Mucor miehei immobilized on a macroporous anionic exchange resin of the phenolic type. Using a 32 factorial experimental design, the effects of temperature within the range of 3545 C and solvent ratios of acetone:isooctane between 0.25:0.75 and 0.75:0.25 (v/v) were evaluated on the activity of the lipase. Lipase from M. miehei revealed a higher residual activity (91%) following a 24 h incubation with the solvent acetone:isooctane at a ratio of 0.25:0.75 (v/v) at 35 C while C. rugosa lipase reached a maximum residual activity of approximately 56% after a 24 h incubation with a solvent acetone:isooctane ratio of 0.25:0.75 (v/v) between 35 and 42 C. For the M. miehei lipase, these results were evaluated experimentally by testing glycerolysis of triolein (biocatalyst initial water activity (aw ) 0.534, molar ratio glycerin:triolein 3:1, amount of protein 90 mg, 24 h). Using the best (35 C, 0.75 Ac) and the worst (45 C, 0.75 Ac) conditions for residual activity in stability assays, it was conrmed that when the predicted optimum conditions were applied, a monoolein yield of over 68% and a total conversion of triolein of approximately 89% were reached. 2011 Elsevier B.V. All rights reserved.

Article history: Received 24 January 2011 Received in revised form 12 April 2011 Accepted 21 April 2011 Available online 30 April 2011 Keywords: Monoglycerides Immobilized lipase Organic medium Response surface methodology

1. Introduction Monoglycerides (MGs) are the most widely used emulsiers in the food, pharmaceutical and other industries. MGs are also considered to be potentially health-benecial emulsiers [1]. Currently, MGs are obtained at industrial scale using nonselective inorganic catalysts at high temperatures (200250 C), starting with fatty acids, followed by direct esterication or reaction with pure triglycerides and by interesterication with glycerine [3]. These chemically catalyzed processes have several drawbacks, such as the generation of dark colored by-products with an undesirable avor and low yields (3040% MGs); therefore, to concentrate

Abbreviations: aw , water activity; MGs, monoglycerides; DGs, diglycerides; P, octanolwater partition coefcient; BSA, bovine serum albumin; pNPP, pnitrophenyl palmitate; pNP, p-nitrophenol; RSM, response surface methodology; RA, residual activity; T, temperature; Ac, acetone volume fraction; R2 , determination 2 , adjusted R2 . coefcient (quadratic correlation coefcient); Radj Corresponding author. Tel.: +57 68879400x55831; fax: +57 68879400x55880. E-mail address: [email protected] (C.E. Orrego). 1381-1177/$ see front matter 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.molcatb.2011.04.017

MGs up to more than 90%, purication by molecular distillation is usually required [2,4]. Using lipases instead of inorganic catalysts to produce MGs and DGs prevents side product formation, can be accomplished at lower temperatures, saves energy and generates light-colored by-products [5,6]. To make this process nancially competitive, it is necessary to immobilize the lipase for re-use [3]. The synthesis of MGs by direct lipase-catalyzed esterication between glycerol and a fatty acid in organic solvents, which improves the poor solubility of the substrates in water, has been proposed in various studies [27]. Bellot et al. [7] investigated MGs synthesis by Rhizomucor miehei lipase via direct esterication between glycerol and oleic acid in organic solvents and proposed that an increase in solvent polarity using mixtures of two solvents drastically improves the selectivity toward MGs formation. Li and Ward proposed an enzymatic method for synthesis of MGs from 1,2-isopropylidene glycerol and n-3 polyunsaturated fatty acid concentrate using lipase IM-60 from Mucor miehei in isooctane and hexane as organic solvents [8]. Kaewthong and H-Kittikun studied the effects of various organic solvents on MG production from theglycerolysis of palm olein with immobilized Pseudomonas sp. lipase.

14

D.M. Cetina et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 1319

The optimal condition for MG production was determined to be in acetone/isooctane mixtures [4]. A useful parameter to characterize solvent properties in relation to lipids and lipases is the octanolwater partition coefcient (P). Log P tends to be largest for compounds with extended non polar structures and smallest or negative for compounds with highly polar groups [9]. Damstrup et al. screened solvent systems of varying polarities for efcient and practical enzymatic glycerolysis of sunower oil using Novozym 435 lipase. They found the maximum MGs production in solvents systems with log P values in the range of 0.31.0 [10]. In non-aqueous media, biocatalysis loss of activity can be caused by deactivation of the enzyme overtime as a consequence of thermal and solvent effects and inhibition of reagents or products. The general method to determine the optimum reaction conditions involves experimental design sets, which are sometimes very complex, in which different solvents, reaction temperatures and reagent molar ratios are tested. The previous study of the non-aqueous residual activity evolution on different solvent and temperature conditions could be a useful approach to identify ranges of probable best (or worst) enzyme behavior. The aim of this work was to study the thermal stability of two immobilized lipases in different mixtures of organic solvents. Two binary mixtures of solvents were used: one with a high log P value (isooctane) and another with a high polarity (acetone). In this way, it was possible to simultaneously screen the inuence of the hydrophobichydrophilic ratios of the reaction media and the temperature of the reaction on the progress of residual catalytic activity. The accuracy of this procedure in dening the enzyme performance on the immobilized lipase systems was veried by experimentally testing the glycerolysis of triolein reaction to form MGs. 2. Materials and methods 2.1. Materials 2.1.1. Enzymes Lipase from Candida rugosa (Lipase AY) was supplied by SigmaAldrich, USA, and lipase from Mucor miehei (Lipozyme ) immobilized on macroporous anion-exchange resin of the phenolic type was purchased from Fluka Analytical, USA. 2.1.2. Immobilization support Chitosan akes (high molecular weight 602 kDa, degree of deacetylation 76.5%) were obtained from SigmaAldrich, USA. 2.1.3. Reagents Glycerol (99%), glycerol trioleate (practical grade 65%), 4nytrophenol and 4-nytrophenyl palmitate (spectrophotometric grade), triolein, monoolein, diolein and oleic acid (analytical grade 99%) were purchased from SigmaAldrich, USA. All other chemicals were of analytical grade and obtained from various sources. 2.2. Methods 2.2.1. Immobilization The procedure for chitosan support production was described in previous works [11,12]. Briey, chitosan akes (4.5 g) were diluted in 150 mL of 1% (v/v) acetic acid solution and treated with glutaraldehyde until the crosslinker content was 0.0003 M. The lms were cured using 30 mL of NaOH/ethanol gelling solution for 2 h. After rinsing with distilled water, gelled matrices were frozen at 45 C for 30 min and thawed for 3 h at 4 C. This freezing and thawing (F/T) step was carried out six consecutive times. The resulting lms were cut into 1 cm 1 cm sections. Lipase from C. rugosa

was immobilized onto chitosan lm supports according to the protocol from Orrego et al. [13]. The enzyme was dissolved and pre-incubated at 35 C in 50 mL of phosphate buffer (pH 7.2) under gentle stirring for 2 h. The chitosan lms were then submerged into the enzyme solution for 20 h at 20 C under agitation (120 rpm). The resultant immobilized lipase on chitosan lms were removed and stored at 4 C. 2.2.2. Protein loading assay The amount of immobilized enzyme on the membrane was determined by measuring the initial and nal concentrations of protein within the enzyme solution used for the immobilization and the washing solutions, using the Biuret method at a wavelength of 540 nm using a spectrophotometer (Model 6405, Jenway, England) [14]. The protein load of lipase was expressed as the percentage of protein (%) on the chitosan membrane. 2.2.3. Water activity pre-equilibration The initial water activity (aw ) values of immobilized enzymes were adjusted by following the gravimetric static method of equilibration as follows: Candida rugosa and Mucor miehei lipases were separately incubated in a chamber containing a saturated salt solution (NaBr), and the system was allowed to reach equilibrium at 40 C for aw = 0.53. The initial water activity of the pre-equilibrated immobilized enzyme system was corroborated using a water activity meter (Decagon, model 99163, Pawkit, USA). 2.2.4. Measurement of lipase synthetic activity in organic media Determination of lipase synthetic activity was based on a lipasecatalyzed transesterication reaction between p-nitrophenyl palmitate (pNPP) and ethanol in n-heptane [15]. The yield of p-nitrophenol (pNP) was determined by a spectrophotometric method (Model 6405, Jenway, England) at a wavelength of 410 nm. The lipase synthetic activity was expressed as mmol of pNP liberated per minute and gram of protein. 2.2.5. Enzymatic stability For lipase-catalyzed monoglyceride production, Damstrup et al. found that a maximum yield for solvents was accomplished when log P values ranged between 0.3 and1.0. In this work, we chose a broader range of log P values (0.462.85) using acetone and isooctane mixtures. The log P value of the mixture of solvents was calculated using a semiempirical formula for related-solvent engineering approaches as suggested by Hilhorst et al. [1619]: log PMix = XAcetone log PAcetone + XIsooctane log PIsooctane (1)

where Xi = component molar fraction, i = acetone or isooctane. Initial and nal enzymatic activities of immobilized lipases were measured before and after Candida rugosa and Mucor miehei lipases were incubated for 24 h in three different concentrations (v/v) of acetone:isooctane (0.25:0.75, log PMix = 2.85; 0.5:0.5, log PMix = 1.43; 0.75:0.25, log PMix = 0.46) at different temperatures (35, 40, 45 C). 2.2.6. Experimental design Response surface methodology (RSM) was used to model the residual activity (RA) of the immobilized enzymes (Candida Rugosa lipase on chitosan and Mucor miehei lipase on macroporous resin). For each lipase assayed, a total of 11 experiments to assess enzymatic stability were carried out following a central composite face centered design as a function of both the temperature (T) and the acetone fraction volume (Ac) in the reaction mixture of acetone and isooctane. The levels considered in both assessed variables are shown in Table 1.

D.M. Cetina et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 1319 Table 1 Codied levels and real values of the experimental factors used in experimental design. Codied levels 1 T ( C) Ac (v) 35 0.25 0 40 0.50 +1 45 0.75

15

2.2.7. Statistical analysis The results of the experimental design were analyzed using Design Expert software version 8 from State-Easy Inc., USA. The linear and quadratic effects of T and Ac and the linear interaction (T) (Ac) on the enzymatic stability of lipases were calculated, and their signicance was evaluated by analysis of variance. A 3D surface, as described by a second order polynomial equation, was tted to each set of experimental data points of residual activity. First and second order coefcients were generated by regression analysis. 2.2.8. Glycerolysis reaction Glycerolysis reactions were performed in 25 mL screw cap asks placed in a thermostated bath (Model 2870, Thermo Electron Corporation, USA) at the desired temperature with shaking at 200 rpm. The reaction mixtures consisted of triolein (4.85 mL, 0.005 gmol), glycerin (1.09 mL, 0.015 gmol) and fresh catalyst (162.60 mg equivalent to 90.00 mg of protein) in acetone:isooctane media. The molar ratio of glycerin to triolein was 3:1. 2.2.9. Analytical methods The course of glycerolysis was monitored by intermittent sampling (200 L) of each reaction sample over 6 h. The organic solvent was completely evaporated in a thermostated bath at 95 C, and samples were re-diluted in hexane. The samples were analyzed to quantify mono-, di- and tri-olein concentrations using an HPLC system (Hitachi, Tokio) equipped with a C-18 reverse phase LiChroCART column and UV detector (215 nm). The levels of free fatty acids (FFA) were determined using the Lowry and Tinsleys colorimetric method [22]. 3. Results and discussion 3.1. Protein load and protein load efciency Protein load and protein load efciency values obtained for Candida rugosa lipase on chitosan membranes were 12.82 mg g1 chitosan and 22.06%, respectively. These values are similar to those obtained in previous works [12]. The protein load of the Mucor miehei lipase was estimated as the protein content in the commercial lipase measured by Kjeldahl analysis (54.44%). 3.2. RSM experiments and model tting The initial enzymatic activity of the evaluated lipases was measured in triplicate before incubating them with the organic solvents. Mucor miehei lipase had the highest initial enzymatic activity, with a value of 35.96 0.63 mol min1 gprot 1 , while Candida rugosa lipase had an enzymatic activity of 16.43 0.27 mol min1 gprot 1 . Assays were used to determine the enzymatic activity of each lipase after 24 h incubation with the organic solvents at a xed temperature. The residual activity (RA) was dened as follows: Residual activity (%) = nal activity 100 initial activity (2)

Fig. 1. Effect of temperature and acetone fraction on the percentage RA of the Mucor miehei lipase.

Table 2 shows the results of nal and residual activities for both lipases. It should be noted that the coded values of each factor (shown in brackets) corresponds to the real value of the factor level. Data in Table 2 were analyzed by RSM to construct an empirical model for the representation of the residual activity of each lipase in terms of the temperature and the acetone fraction in the reaction medium. Based on regression analysis at a 95% condence interval, a quadratic model was used to t the observed data by least squares analysis, and the signicance of the lack of t error and p-values of the parameter estimations were determined. ANOVA results are shown in Table 3. The F-value of the models were 47.56 and 81.00, with a corresponding p-value of 0.0003 and <0.0001, for Mucor miehei and Candida rugosa lipases, respectively, implying that the model was signicant at the 95% condence level. The effects of the two synthesis variables as well as their interactions can be evaluated based on their p-values. The p-value at which every term in the selected model should be signicant was set to 0.05. While the interaction term T(Ac) was the most signicant factor affecting the percentage of RA of the Mucor miehei lipase, the independent variable temperature (T) and the quadratic term of temperature (T2 ) also inuenced the response. The independent variable and the quadratic term of the acetone fraction in the reaction medium did not cause a significant effect on the percentage of RA within the designed intervals. The residual activity of the Candida rugosa lipase was more greatly inuenced by the independent variable acetone fraction (Ac) than that of the Mucor miehei. The independent variable acetone fraction (Ac) and the quadratic terms of the temperature and the acetone fraction (T2 and Ac2 ) also signicantly affected the residual activity of Candida rugosa, while the response for the Candida rugosa lipase was not affected by the interaction term T(Ac). The empirical models were obtained for predicting RA for both 2 show a close agreelipases (Table 4). The high values of R2 and Radj ment between the experimental results and the theoretical values predicted by the models. The effects of temperature (T) and acetone fraction (Ac) on the residual activity of the Mucor miehei and Candida rugosa lipases are shown in Figs. 1 and 2, respectively. Catalytic activity and stability in non-aqueous media are directly correlated with solvent hydrophobicity (high log P) because nonpolar and non-water miscible media, as opposed to polar and miscible media, do not withdraw the bound water crucial for the enzyme structure and activity [23]. In some cases, the addition of small amounts of water-miscible solvents such as acetone results in enhanced enzyme activity

16

D.M. Cetina et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 1319

Table 2 Experimental conditions of central composite design (CCD) runs of Design-Expert 8.0 and corresponding results (the response). Run Level variables T ( C)

Response value Ac (v)

Final activity (mol min1 gprot 1 ) M.Mb C.Rc 9.05 7.64 4.34 3.80 5.29 5.23 5.27 5.47 3.96 9.26 3.96

Residual activity (%) M.Mb 47.28 78.33 91.06 45.54 78.99 79.53 78.36 73.36 55.85 79.16 79.29 C.Rc 55.06 46.52 26.40 23.11 32.26 31.89 32.13 33.35 24.15 56.46 24.15

1 2 3 4 5 6 7 8 9 10 11
a b c

35 (1) 45 (1) 35 (1) 45 (1) 40 (0) 40 (0) 40 (0) 35 (1) 45 (1) 40 (0) 40 (0)

0.25 (1) 0.25 (1) 0.75 (1) 0.75 (1) 0.5 (0) 0.5 (0) 0.5 (0) 0.5 (0) 0.5 (0) 0.25 (1) 0.75 (1)

17.01 28.17 32.69 16.35 28.36 28.55 28.13 26.38 20.09 28.47 28.52

Acetone fraction (volume). Mucor miehei lipase. Candida rugosa lipase immobilized in chitosan.

Table 3 ANOVA results for the selected models of Design-Expert 8.0 for the response of each lipase. Lipase Mucor miehei Source Model T Ac T (Ac) T2 (Ac)2 Lack of t Pure error Model T Ac T (Ac) T2 (Ac)2 Lack of t Pure error Sum of squares 2179.15 170.49 20.61 1465.46 496.26 0.99 45.13 0.69 1451.61 73.76 1186.95 6.87 23.93 182.23 17.59 0.33 Degree of freedom 5 1 1 1 1 1 3 2 5 1 1 1 1 1 3 2 Mean square 435.83 170.49 2.25 1465.46 496.26 0.99 15.04 0.34 290.32 73.76 1186.95 6.87 23.93 182.23 5.86 0.16 F-Value 47.56 18.61 20.61 159.93 54.16 0.11 43.90 81.00 20.58 331.14 1.92 6.68 50.84 35.59 p-Value Prob > F 0.0003 0.0076 0.1940 <0.0001 0.0007 0.7558 0.0224 <0.0001 0.0062 <0.0001 0.2247 0.0492 0.0008 0.0275

Candida rugosa

Fig. 2. Effect of temperature and acetone fraction on the percentage RA of the Candida rugosa lipase.

and stability [24]; however, when the concentration is increased, most water-miscible solvents have an inhibitory effect on the biocatalyst mainly due to changes in the afnity of the enzyme for the substrate [25]. In the present work, the Candida rugosa lipase immobilized on chitosan displayed this behavior. There was a noticeable effect of increasing the acetone concentration on the residual activity (RA), which was reduced to 23.11% when the volume fraction of acetone in reaction media was 0.75. In contrast, a maximum residual activity of 56.46% was found for the 0.25 acetone volume fraction. For this enzyme, the modeled RA was inuenced by both temperature and solvent concentration, with the latter having the primary effect on the response (p < 0.0001). The model also predicted a more favorable outcome for RA within the temperature range 3540 C. Similar results for RA were obtained in a study with Candida Rugosa lipase immobilized on chitosan for esterication in hexane. This catalyst had an RA up to 80% after 12 h incubation in the solvent [26]. The performance with respect to RA for immobilized Mucor miehei lipase showed contrary results. In this case, the most advantageous combination of temperature and solvent mixture was a

Table 4 Model equations for the response surfaces tted to the experimental data points from percentage of residual activity, as a function of the temperature (T) and acetone fraction 2 . (Ac) of Mucor miehei and Candida rugosa lipases, and respective R2 and Radj Responses Residual activity M.M (%) Residual activity C.R (%) Model equations RAM.M = 1081.75 + 51.38T 15.31T(Ac) 0.56T2 RAC.R = 53.44 + 8.61T 233.91Ac 0.13T2 + 135.70(Ac)2 R2 0.98 0.99
2 Radj

0.96 0.97

D.M. Cetina et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 1319

17

low temperature (35 C) and a high acetone concentration (0.75low log P), with 91% of RA after 24 h. In a comparative study on the effect of solvent media on Candida antarctica lipase-catalyzed glycerolysis, superior monoglyceride production was found to occur in high content acetone solvent [27]; however, in this study, the Mucor miehei lipase demonstrated high RA performance (84% of RA) at 4243 C and a 0.25 volume fraction of acetone in the solvent mixture. It should be noted that the immobilized Mucor miehei lipase in the solvent systems analyzed in this work is an example of an enzymatic reaction system for which there is no correlation between activity retention and the log P value as cited above. The effects of organic solvents on enzyme activity are primarily due to the impact of the solvent on interactions with the essential enzymebound layer of water rather than with the enzyme itself [28]. In non-aqueous catalysis, immobilized enzymes may compete not only with the solvent but also with the support, as preserving their essential monolayer of water on the surface is required for their biocatalytic function. The hydrophilicity/hydrophobicity characteristics of the solvent and the support play important roles in the preservation of the water bound to the enzyme. In some cases, the type of catalyzed reaction also could affect the availability of water to the enzyme. A total loss of activity for Lipozyme (the same lipase used in this work) catalyzed esterication was detected after its rst use. The water produced by the reaction was accumulated on the support and facilitated the enzyme denaturalization [29]. In this type of system, it was suggested that the presence of a polar solvent in the reaction media could remove the water overload and consequently preserve the enzyme in its active state [30,31]. Chitosan has higher water capacity (aquaphilicity) than the Lipozyme support. The presence of acetone in the reaction system could congure a water-competitive environment in which an enzyme (in the present work, Candida rugosa lipase) is unable to preserve its bound water. This interaction with the available water could explain the poor results of this biocatalyst in media containing a high concentration of acetone solvent. 3.3. Glycerolysis of triolein Due to the superior residual activity (RA) of Mucor miehei lipase in most of the assays after 24 h incubation, the glycerolysis reaction was conducted using this biocatalyst. There were two zones of high RA observedfor Mucor miehei lipase: 35 C/0.75 Ac and 45 C/0.25 Ac, with 91% and 84% of RA, respectively. There were also two zones of low RA observed for the same lipase: 35 C/0.25 Ac and 45 C/0.75 Ac, with 47% and 45% of RA, respectively. In this work, a nonaqueous enzymatic activity method based in a transesterication reaction was used for the stability study (instead of the commonly used measurement of lipase activity by hydrolysis). Therefore, it is possible that the predictions of the RSM stability analysis can also forecast the catalytic behavior of the enzyme in the glycerolysis reaction in organic medium. In order to test this hypothesis, the glycerolysis of triolein reaction was conducted in duplicate in the above mentioned four zones: two for high RA and two for low RA. The initial water activity of the biocatalyst was controlled to 0.53, the molar ratio of glycerin:triolein was 3:1 and the amount of enzyme in the reaction media was 90 mg. Figs. 36 show the reaction evolution. Monoolein, diolein, and oleic acid production and non-reacted triolein concentration were plotted against time for two reaction trials. The ability to superimpose these graphs demonstrates the reproducibility of these assays. Fig. 3 shows the glycerolysis of triolein in the rst zone of high RA (35 C, 0.75 Ac). After 24 h of reaction, the mass compositions of monoolein, diolein and oleic acid were 68.39%, 17.53% and 3.52%, respectively. The conversion of triolein was approximately 89.04%. The effects seen in these results are explained by the polarity and

100 90 80 Triolein Monolein Diolein Oleic Acid

Composition (% mass)

70 60 50 40 30 20 10 0 0 5 10

15

20

25

Time (hours)
Fig. 3. Glycerolysis of triolein catalyzed by Mucor miehei lipase at 35 C in acetone:isooctane (0.75:0.25, v/v). High RA zone. First () and second () assays.

100 90 80 Triolein Monolein Diolein Oleic Acid

Composition (% mass)

70 60 50 40 30 20 10 0 0 5 10 15 20

25

Time (hours)
Fig. 4. Glycerolysis of triolein catalyzed by Mucor miehei lipase at 45 C in acetone:isooctane (0.25:0.75, v/v). High RA zone. First () and second () assays.

the hydrophobicity of the organic solvents on the reaction, as it has been shown that increasing the proportion of the polar solvent shifts the equilibrium toward MG production. In fact, the transesterication of ethyl oleate with glycerol at optimized conditions up to 68% revealed that MG were formed after 12 h by lipase from Candida antarctica in acetone [32]. In other published works studying optimization of monoolein production [2,4,6,33], lower monoolein yields (20.1254.38%) were reported. Fig. 4 shows the evolution of the reaction in the second zone of high RA (45 C, 0.25 Ac). The production of monoolein, diolein

100 90 80 Triolein Monolein Oleic Acid Diolein

Composition (% mass)

70 60 50 40 30 20 10 0 0 5 10 15 20

25

Time (hours)
Fig. 5. Glycerolysis of triolein catalyzed by Mucor miehei lipase at 35 C in acetone:isooctane (0.25:0.75, v/v). Low RA zone. First () and second () assays.

18
100 90 80

D.M. Cetina et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 1319

Triolein Monolein Diolein Oleic Acid

70 60 50 40 30 20 10 0 0 5 10 15 20

ture (35 C) and the highest fraction of acetone in the solvent system (0.75). Conversely, for the Candida rugosa lipase, the maximum RA (56%) was achieved at the lowest acetone concentration in reaction system (0.25) and at approximately 36 C. Surface response modeling proved to be valuable. The equations developed can be predictive of the effects of solvent ratio and temperature on the lipase RA selected response. The use of the RSM-correctly predicted higher and lower RA log P/temperature conditions for the superior and inferior conversions of triolein in the glycerolysis trials. These results suggest the effectiveness of this procedure in the determination of the performance of immobilized lipase systems in non-aqueous media lipase catalysis optimization.
25

Composition (% mass)

Time (hours)
Fig. 6. Glycerolysis of triolein catalyzed by Mucor miehei lipase at 45 C in acetone:isooctane (0.75:0.25, v/v). Low RA zone. First () and second () assays.

Acknowledgements The authors are grateful to the Universidad Nacional de Colombia for nancial support and to COLCIENCIAS and its program Jvenes Investigadores e Innovadores 2008. References
[1] L. Sagalowicz, M.E. Leser, H.J. Watzke, M. Michel, Trends Food Sci. Technol. 17 (2006) 204214. [2] W. Kaewthong, S. Sirisansaneeyakul, P. Prasertsan, A. H-Kittikun, Process Biochem. 40 (2005) 15251530. [3] S. Ferreira-Dias, A.C. Correia, M.M.R. Fonseca, J. Mol. Catal. B: Enzym. 21 (2003) 7180. [4] W. Kaewthong, A. H-Kittikun, Enzyme Microb. Technol. 35 (2004) 218222. [5] Y. Yesiloglu, JAOCS 81 (2004) 157160. [6] S. Ferreira-Dias, A.C. Correia, F.O. Baptista, M.M.R.d. Fonseca, J. Mol. Catal. B: Enzym. 11 (2001) 699711. [7] J.C. Bellot, L. Choisnard, E. Castillo, A. Marty, Enzym. Microb. Technol. 28 (2001) 362369. [8] Z.-Y. Li, O.P. Ward, JAOCS 70 (1993) 745748. [9] D.R. Lide, in: C.P.D.R. Lide (Ed.), OctanolWater Partition Coefcients, Handbook of Chemistry and Physics, Boca Raton, FL, 20042005. [10] M.L. Damstrup, T. Jensena, F.V. Sparsc, S.Z. Kiilb, A.D. Jensenb, X. Xua, JAOCS 82 (2005) 559564. [11] C.E. Orrego, J.S. Valencia, Bioprocess Biosyst. Eng. 32 (2009) 197206. [12] C.E. Orrego, J.S. Valencia, C. Zapata, Catal. Lett. 129 (2009) 312322. [13] C.E. Orrego, N. Salgado, J.S. Valencia, G.I. Giraldo, O.H. Giraldo, C.A. Cardona, Carbohyd. Polym. 79 (2010) 916. [14] R.M. Robson, D.E. Goll, M.J. Temple, Anal. Biochem. 24 (1968) 339341. [15] Y. Teng, Y. Xu, Anal. Biochem. 363 (2007) 297299. [16] C. Laane, S. Boeren, K. Vos, C. Veeger, Biotechnol. Bioeng. 30 (1987) 8187. [17] W. Du, D. Liu, L. Dai, Biotechnol. Progr. 23 (2007) 10871090. [18] Z. Guo, Y. Sun, Biotechnol. Progr. 20 (2008) 500506. [19] R. Hilhorst, R. Spruijt, C. Laane, C. Veeger, Eur. J. Biochem. 144 (1984) 459466. [22] R.R. Lowry, I.J. Tinsley, JAOCS 53 (1976) 470472. [23] A.R.M. Yahya, W.A. Anderson, M. Moo-Young, Enzym. Microb. Technol. 23 (1998) 438450. [24] N. Nawani, R. Singh, J. Kaur, Electron. J. Biotechnol. 9 (2006) 559565. [25] M.H. Vermue, J. Tramper, Pure Appl. Chem. 67 (1995) 345373. [26] E.B. Pereira, G.M. Zanin, H.F. Castro, Braz. J. Chem. Eng. 20 (2003) 343355. [27] Y.-C. Yang, S.R. Vali, Y.-H. Ju, J. Chin. Inst. Chem. Eng. 34 (2003) 617623. [28] G. Carrea, S. Riva, Angew. Chem. Int. Ed. 39 (2000) 22262254. [29] G.B. Oguntimein, W.A. Anderson, M. Moo-Young, Biotechnol. Lett. 17 (1995) 7782. [30] C.J. Gray, J.S. Narang, S.A. Barker, Enzym. Microb. Technol. 12 (1990) 800 807. [31] M.M. Hoq, T. Yamane, S. Shimizu, JAOCS 62 (1985) 10161021. [32] U.T. Bornscheuer, Enzym. Microb. Technol. 17 (1995) 578586. [33] H.A. Aksoy, M. Tuter, Chem. Eng. Commun. 92 (2005) 1417. Diana Marcela Cetina received her B.S. and M.Sc. degrees in Chemical Engineering at Universidad Nacional de Colombia. Her research interests include non aqueous enzyme catalysis and oleochemistry.

and oleic acid after 24 h of reaction was approximately 31.64%, 41.02% and 13.99%, respectively. The total conversion of triolein was 85.25%. In this case, despite the acceptable value of substrate conversion of the reaction, the monoolein yield was poor, and diolein was the main reaction product. This may be because the production of diolein is favored by the lower polarity of the reaction medium. These results agree with another study in which higher diglyceride synthesis took place in medium rich in non-polar organic solvents [32]. In another published work, in the direct esterication between glycerol and oleic acid catalyzed by Rhizomucor miehei lipase, the diolein production increased as the proportion of 2-methyl-2-butanol (polar solvent) in n-hexane decreased [7]. Fig. 5 shows the product compositions of the reaction in one zone of low RA (35 C, 0.25 Ac). The compositions of monoolein, diolein and oleic acid after 24 h of reaction were approximately 23.98%, 16.82% and 12.98%, respectively, while the total conversion of triolein was 52.85%. These conversions to triolein and monoolein were considered low. Almost identical results were found in the second zone of low RA (45 C, 0.75 Ac), shown in Fig. 6, as after 24 h of reaction, the compositions of monoolein, diolein and oleic acid were approximately 28.12%, 18.72% and 5.64%, respectively, and the total conversion of triolein was 51.42%. The main difference between these assays was the diolein yield prole, which was bell shaped for the rst low RA zone assay and an approximately hyperbolic curve for the second assay. In conclusion, the results obtained by the glycerolysis of triolein assays were consistent with the behavior predicted trough the preliminary experimental setup for lipase stability. In the two zones in which a high enzyme RA was predicted, the conversion of triolein was also increased with a monoolein yield considerably higher at 35 C and 0.75 volume fraction of acetone. Conversely, in the two solvent/temperature zones that previously showed poor stability (low RA), the nal compositions and the conversion for the glycerolysis reaction were very similar, and in both cases, the monoolein yield was poor. 4. Conclusions For the selection of a suitable lipase catalytic system for monoglyceride production, we studied the catalytic residual activity of two immobilized lipases (Candida rugosa on chitosan and Mucor miehei on macroporous anion-exchange resin of the phenolic type) in three different mixtures of acetone/isooctane at different temperatures over 24 h. The conditions that optimized the residual activity RA (%) of the Mucor miehei enzyme were the lowest assayed reaction tempera-

D.M. Cetina et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 1319

19

Gloria Ins Giraldo received the M.Sc. and Ph.D. degrees in Sciences-Chemistry at Universidad Nacional de Colombia. She is currently associate professor and Vice Dean of Faculty of Sciences at the Universidad Nacional de Colombia of Manizales. Her research areas are protein stability, enzymatic catalysis, physical and chemical characterization of foods and biological materials, and food preservation.

Carlos Eduardo Orrego is currently Professor and the Head of the Instituto de Biotecnologa y Agroindustria at the Universidad Nacional de Colombia of Manizales. He received his Ph.D. degree at Universidad Nacional de Colombia in Sciences-Chemistry. His research interests include biofuels, industrial biotechnology, enzyme catalysis and freeze-drying and spray drying of food and non food materials.