Evaluation of the performance of differently immobilized recombinant lipase B from Candida antarctica preparations for the synthesis of pharmacological derivatives in organic media

Abstract

This paper shows the production of lipase B from Candida antarctica (LIPB) after cloning the gene that encoded it in Pichia pastoris using PGK as a constitutive promoter. The production of the lipase is lower using this strategy but it avoids the use of inducers like methanol. The performance of this enzyme was compared with that of the commercial enzyme (CALB) after immobilization on different supports in different reactions. As supports, we used Accurel 1000, and three core–shell supports (poly(methyl methacrylate) on the core and on the shell – PMMA/PMMA; poly(methyl methacrylate-co-divinylbenzene) on the core and on the shell – PMMA-co-DVB/PMMA-co-DVB; and poly(styrene-co-divinylbenzene) on the core and on the shell – PS-co-DVB/PS-co-DVB). The popular Novozym 435 was also utilized to assess the features of the new biocatalysts. All these supports adsorbed lipases via interfacial activation of the open form of the lipase on the hydrophobic surface of the supports. The studied reactions were esterification of oleic acid and ethanol in a solvent-free medium, resolution of (±)-1,3,5-O-benzyl-myo-inositol via acylation using vinyl acetate in hexane and resolution of (±)-1,2-O-isopropylidene-3,6-di-O-benzyl-myo-inositol via acylation using vinyl acetate (solvent free system). The results varied depending on the employed supports and on the studied reactions, but some general trends may be observed, pointing to better behavior of LIPB compared to CALB. The use of 4 different supports gave more strength to these differences, as it did not depend on a specific difference between a single support/enzyme pair, but it is more general. Thus, LIPB seems to have some advantages compared to the commercial enzyme on all the reactions assayed in this paper. PS-co-DVB/PS-co-DVB-LIPB is in general the most active preparation (even 50% higher activity was observed). Further investigations are in development to determine the structural reasons for these differences.

---

1. Introduction

Lipases are often used in biotechnology, currently accounting for around 5% of the global enzyme market. Lipase B from Candida antarctica (CALB) is one of the most widely used lipases in biocatalysis.^1,2 It is highly regio- and enantioselective, and has been used in the synthesis of esters and enantioselective resolution of a wide amount of compounds, like amines, acids, and mainly alcohols.^3–11

Despite the great industrial importance of CALB – it is already sold in free and immobilized form¹² by different companies (e.g., Novozym 435 from Novozymes¹³) – great effort is being put into finding other CALB biocatalysts that are even more efficient. The efforts devoted to improve this lipase utility in the industry should involve all the production of the biocatalysts, from improvements on enzyme production to improvements on immobilization process.

In this context, Pichia pastoris is one of the most widely used yeasts to express heterologous proteins because of its secretory capacity and ability to mass-produce industrial enzymes, including lipases.¹⁴ The production method for biomolecules secreted by Pichia pastoris must consider the promoter which the protein gene is linked to. The most studied promoter is AOX.^14,15 In this case, the gene that is introduced needs to be induced by methanol and the induction step has to be preceded by a biomass growth phase. However, the different steps involved and the use of methanol make the scale-up process more complex and, in some cases, even unfeasible, depending on the product.

In order to overcome the barriers inherent to use the AOX promoter, constitutive expression promoters have been widely investigated, since it is possible to obtain a culture associated with growth without needing to add neither methanol nor any other inducer. The promoter of the glyceraldehyde-3-phosphate dehydrogenase gene (pGAP) is the most widely used constitutive expression promoter because it is involved in the glycolysis, which is well regulated.¹⁶ However, there are other constitutive promoters, like PGK (from the enzyme phosphoglycerate kinase I), whose regulation is also associated with the glycolytic pathway, and which has been little explored in the literature.^17,18

Another point to be considered to have a more fruitful design of the biocatalysts is the culture medium. Glycerin is a by-product in the production of biodiesel, and has become almost a residue.¹⁹ Thus, currently there are many efforts to find uses to glycerin, and in the case of our proposal, it may be interesting to evaluate their use as carbon source in the production of this new recombinant CALB.

Once the enzyme has been produced in a soluble form in the culture medium, purification is required to ensure the reaction specificity by eliminating other esterases in the crude extract, and, in some cases, the immobilization process is needed. A proper immobilization protocol should facilitate the enzyme reusability and, in certain cases, its stability and efficiency.^20–26 In an ideal situation, the coupling of immobilization and purification will save time and costs.²⁷ In the case of lipases, this can be achieved by using hydrophobic supports to immobilize lipases at low ionic strength²⁸ immobilization strategy that even greatly improves the enzyme stability.²⁹ This immobilization technique involves the interfacial activation of the lipase versus the hydrophobic surface of the support,³⁰ and this is the main cause of the lipase stabilization and improved activity. In homogenous aqueous medium, the active center of the lipases is usually secluded from the reaction media by a polypeptide chain, called lid (closed form).^31–34 Depending on the external conditions, this lid may move exposing to the medium a large hydrophobic pocket, formed by the internal face of the lid and the areas surrounding the lipase active center. This form of the enzyme, called open form, allows its readily adsorption on any hydrophobic surface, such as oil droplets,^35,36 hydrophobic supports,³⁷ open forms of other lipase molecules (homogeneous or heterogeneous dimmers, which may be formed),^38,39 other hydrophobic proteins.⁴⁰

Therefore, as well as the above mentioned strategies, mastery of technologies for synthesizing tailor-made supports could be great advantage point to obtain an even more efficient biocatalyst.^41–44 In this context, hydrophobic core–shell polymeric supports synthesized by a combined suspension and emulsion polymerization process have shown a great potential on lipases immobilization procedures.^5,45–48 This technique has already been described in the literature.^45–48 It involves two fundamental reaction steps. Firstly, a standard suspension polymerization process is carried out in order to produce the polymeric cores. Then, on a second step, the emulsion constituents are added to the reaction system, resulting in the synthesis of in situ nanoparticles. Therefore, these nanoparticles coagulate over the previously produced cores to form a porous shell, resulting in porous core–shell polymeric particles. The final morphology of the particles is affected by the reaction conditions and the compounds added in each step.^45–48

The versatility of these supports may be used to tune enzyme properties and permit the design of these supports, allowing the development of new biocatalysts for specific reaction media.^45–48

Therefore, the aim of this study was to express recombinant lipase B from Candida antarctica (we will call this lipase as LIPB to distinguish this enzyme from the commercial CALB) in Pichia pastoris using PGK as a constitutive promoter.⁴⁹ In this work LIPB and CALB were also immobilized on the home-made core–shell polymeric particles, and the performance of the new biocatalysts was compared to that of the commercial one in kinetic resolutions of different pharmaceuticals using different strategies.

2. Materials and methods

2.1. Materials

Solutions of free Candida antarctica lipase B (CALB) (19.11 mg protein per mL) and immobilized Novozyme 435® were kindly donated by Novozymes (Spain). Accurel MP 1000 was purchased from Membrane GmbH (Germany). Styrene (S – Nitriflex Resinas S/A, minimum purity of 99.9 wt%), methyl methacrylate (MMA – Aldrich, minimum purity of 99.9) and the crosslinking agent, divinylbenzene (DVB – Aldrich, minimum purity of 99 wt%), were employed in the polymerization reactions. (±)-1,3,4-Tri-O-benzyl-myo-inositol (DL-1) and (±)-1,2-O-isopropylidene-3,6-di-O-benzyl-myo-inositol (DL-2) were obtained from Laboratory Roderick Barnes, UFRJ, RJ, Brazil. Bradford's reagent was obtained from BIO-RAD. All other reagents and solvents were of analytical grade.

2.2. Methods

All the experiments were performed in triplicate (n = 3) and the values that are shown correspond to their means associated with their corresponding standard deviation.

2.2.1. Recombinant strain of Pichia pastoris. The X-33 wild strain of Pichia pastoris was transformed previously by inserting a constitutive expression vector pPGKΔ3_PRO_LIPB, constructed from a synthetic gene of Candida antarctica lipase B (sequencing LIPB). The cells of transformed P. pastoris were grown in a medium composed of 1% (w/v) yeast extract, 2% (w/v) peptone and 2% (w/v) glycerol (YPG), stored in 25% glycerol at −80 °C.⁴⁹

2.2.2. Production of LIPB expressed in P. pastoris. The cells were reactivated on the same aforementioned media (see Section 2.2.1) in Erlenmeyer flasks with 200 mL volume, incubated at 30 °C, stirred at 200 rpm, for 20 h. The inoculum on the bioreactor (New Brunswick Scientific – Bioflo 310) was made using 1 g L⁻¹ of biomass, the used vase present 3 L of working volume. The cultivation medium contained: 2.0 g L⁻¹ citric acid, 12.4 g L⁻¹ (NH₄)₂HPO₄, 0.022 g CaCl₂·2H₂O, 0.9 g L⁻¹ KCl, 0.5 g MgSO₄·7H₂O, 100 g L⁻¹ glycerol, and 4.6 mL L⁻¹ PTM1 trace salts stock solution. PTM1 solution contained: 6.0 g L⁻¹ CuSO₄·5H₂O, 0.08 g L⁻¹ NaI, 3.0 g L⁻¹ MnSO₄·H₂O, 0.2 g L⁻¹ Na₂MoO₄·2H₂O, 0.02 g L⁻¹ H₃BO₃, 0.5 g L⁻¹ CoCl₂, 20.0 g L⁻¹ ZnCl₂, 65.0 g L⁻¹ FeSO₄·7H₂O, 0.2 g L⁻¹ biotin, and 5.0 mL H₂SO₄ per L solution (95–98%).⁵⁰ The pH was set to 7.0 with 5% H₂SO₄ and 20% NH₄OH.

The temperature, pH, and dissolved oxygen (DO) were controlled throughout the whole cultivation period. Oxygenation was controlled in order to maintain dissolved oxygen saturation at 30%. A cascade control was used, with a stirrer range from 250 to 700 rpm and gas flow of 0–3 L min⁻¹.

Nitrogen, glycerol, protein, cell concentration, and lipase activity were analyzed periodically.

2.2.3. Nitrogen determination. The methodology was adapted from Fawcett and Scott⁵¹ and Tabacco and collaborators.⁵²

2.2.4. Glycerol determination. Glycerol was determined by high-performance liquid chromatography (HPLC, Agilent Technologies) equipped with an HPX-87H column (BioRad – 300 mm × 7.8 mm). The temperature was maintained at 65 °C and a 0.005 M solution of sulfuric acid at 0.6 mL min⁻¹ was used as the mobile phase. The sample volume was 20 μL using IR (Refractive Index) as detector.

2.2.5. Determination of protein. The methodology described by Bradford⁵³ was used.

2.2.6. Determination of cell concentration. Cell concentration was determined in a spectrophotometer (600 nm). The correlation factor between cell absorbance and its concentration was 0.411 g L⁻¹.

2.2.7. Lipase activity.
2.2.7.1. Determination of lipase hydrolytic activity using p-NPL. Hydrolytic activity was monitored spectrophotometrically at 412 nm using 0.25 mM p-nitrophenyl laurate (p-NPL) in 25 mM sodium phosphate/5% (v/v) DMSO/5% acetonitrile (v/v) at pH 7.0 and 30 °C as described by Cunha et al.⁵ For the soluble samples, the reaction was conducted using 0.05 mL of a lipase solution in 2.45 mL a substrate solution.

For immobilized lipases, the reaction was started by the addition of 0.1 g of immobilized enzyme in 9 mL substrate solution in a flask. Periodically samples of 1 mL of the reaction were filtered and the absorbance was determined at 412 nm.

One activity unit (U) was defined as the amount of enzyme necessary to hydrolyze 1 μmol of p-NPL per minute under assay conditions. Enzymatic activity was determined as the average of triplicates.

2.2.7.2. Determination of lipase hydrolytic activity using tributyrin. The hydrolysis of 100 mM tributyrin was performed at 40 °C and pH 7 using a pHstat and a 40 mM NaOH solution was employed as titrating reagent. One unit of lipase activity (U) is defined as the quantity of enzyme necessary to catalyze the production of 1 μmol of butyric acid (supernatant analysis) per minute under the assay conditions.
2.2.7.3. Determination of esterification activity of the immobilized lipase. The esterification activity of the immobilized enzyme was determined as the initial rate in esterification reactions between oleic acid and ethanol at a molar ratio of 1 : 1 at 40 °C using a pHstat, with 5 wt% of enzyme in relation to the substrates followed by Cunha et al.^5,63 After the addition of the enzyme to the substrates, the mixture of oleic acid and ethanol were collected on different times and stopped with acetone : ethanol (1 : 1). The remainder oleic acid content was determined by titration with 0.03 M NaOH. One lipase activity unit (U) was defined as the amount of enzyme necessary to consume 1 μmol of oleic acid per minute.

2.2.8. Preparation of core–shell polymeric supports. As previously mentioned, core–shell polymeric supports were synthesized through a combined suspension/emulsion polymerization process. The reactions were carried out in an open 1 L jacketed glass reactor at 85 °C. Three different supports were produced: core–shell particles based on a copolymer of styrene and divinylbenzene (which corresponded to 25% wt of the monomer mixture) on both core and shell; core–shell particles formed by a copolymer of methyl methacrylate and divinylbenzene (which also corresponded to 25% wt of the monomer mixture) on the core and the shell; core–shell particles based on an homopolymer of methyl methacrylate on the core and the shell.^5,45 Styrene/divinylbenzene supports have been described as very useful supports for lipase immobilization.^5,54–56

Initially, the core particles were formed through standard batch suspension polymerization by dispersing the monomer mixture, containing benzoyl peroxide as an initiator, in an aqueous solution (containing poly(vinyl alcohol) as a stabilizer) under agitation (kept at 950 rpm in all experiments). After two hours of suspension reaction, the emulsion constituents (monomer mixture and aqueous solutions of potassium persulfate, used as an initiator, and lauryl sulfate, used as an emulsifier) were added to the reaction medium. In order to control the reaction temperature, the monomer mixture was added continuously at 0.04 L h⁻¹. After the end of the semi-batch emulsion step, reaction was conducted for two hours to favor the coverage of the core particles. At the end of each experiment, the reactor was cooled down and the particles were filtered, washed several times with distilled water, and dried.^5,45,48 It is possible to observe the reaction scheme on Fig. 1. It is important to mention that, depending on the supports composition, specific functional groups may be found on the particles surface. Divinylbenzene and styrene based particles contain phenyl groups on their surfaces.⁵⁷ On the other hand, methyl methacrylate compounds show ester groups, as functional groups, on their surfaces.⁵⁸ No covalent reaction may be expected, but different interactions are expected to exist between the polymeric particles and the studied enzymes.

Fig. 1 Scheme of the combined suspension–emulsion polymerization reaction.

2.2.9. Characterization of core–shell supports. The morphological properties of the supports (specific area, average pore size) were determined by nitrogen physisorption using a Micromeritics analyzer (model ASAP 2000) applying the standard BET model. Each sample was treated at 60 °C. It is important to mention that the morphology of the core–shell polymeric particles was shown in a previous study.⁵

The features of the synthesized particles are shown in Table 1. The commercial support Accurel MP 1000 (polypropylene) was also used due to the successful performance of this support in lipase immobilization.^30,59–61

Table 1 Morphological properties of the supports^a employed in CALB and LIPB immobilization

Supports	Surface area (m^2 g^−1)	Average pore size (Å)
a These results were shown in Cunha et al.^5b This is a polymer support based on a copolymer of styrene and divinylbenzene in the core and the shell.c This is a polymer support based on a copolymer of methyl methacrylate and divinylbenzene in the core and the shell.d This is a polymer support based on an homopolymer of methyl methacrylate in the core and the shell, respectively, with 2 hours of suspension.^5
Accurel MP 1000	39.0	230.0
PS-co-DVB/PS-co-DVB^b	28.9	190.4
PMMA-co-DVB/PMMA-co-DVB^c	12.6	193.4
PMMA/PMMA^d	5.0	211.6

---

2.2.10. Immobilization of lipases on commercial and core–shell supports. One gram of Accurel MP 1000 or the core–shell supports – PS-co-DVB/PS-co-DVB (styrene copolymerized with divinylbenzene on core and shell), PMMA/PMMA (poly(methyl methacrylate) on core and shell) and PMMA-co-DVB/PMMA-co-DVB (methyl methacrylate copolymerized with divinylbenzene on core and shell) – were added to the corresponding volume of a solution of CALB (commercial lipase) or LIPB (recombinant lipase) in 5 mM sodium phosphate at pH 7 and 25 °C. In order to study the maximum loading capacity of each core–shell support, different units of activity were added to each solution (50, 200, 350 and 700 U of p-NPL hydrolytic activity), as shown in ESI section.†

The activity of the supernatant and suspension was monitored using p-NPL. Immobilization was considered to be complete when no significant changes in the activity of the supernatant were detected after 24 h. After immobilization, the suspension was filtered and the supported enzyme was washed several times with distilled water.

2.2.11. Resolution of racemic myo-inositol derivatives. The enzymatic reactions were conducted in closed reactors with thermostats under the optimal reaction conditions for each substrate previously described:

(±)-1,3,5-O-Benzyl-myo-inositol (DL-1): 8 mg substrate dissolved in 2.0 mL vinyl acetate in hexane (1 : 5.5) and 222.8 U of biocatalyst (80 mg) at 45 °C, during 8 h of reaction time for the core–shell derivatives and Novozyme 435, and 8 h of reaction time (immobilized LIPB and Novozym 435) or 12 h (lyophilized LIPB, solution commercial CALB and immobilized CALB).⁶²

(±)-1,2-O-Isopropylidene-3,6-di-O-benzyl-myo-inositol (DL-2): 10 mg substrate dissolved in 2.0 mL vinyl acetate (solvent-free medium) at 45 °C and 71 U of biocatalyst (26 mg) using p-NPL; when immobilized LIPB and Novozyme 435 were employed, the reactions occurred for 8 h; when lyophilized LIPB, solution commercial CALB and immobilized CALB were used, the reactions occurred for 13 h.⁶³

The reactions were halted by removing the biocatalyst from the reaction medium, and the solvent was evaporated in a vacuum concentrator (Speed Vac Concentrator – Savant SPD 1010-Thermo Scientific). The samples were resolubilized in 1 mL of the mobile phase acetonitrile : H₂O (60 : 40) for the reverse-phase chiral HPLC analysis. The conversion rate and enantiomeric excess of the DL-1 and DL-2 derivatives were calculated as described in ref. 62 and 63, respectively.

2.2.12. Operational stability in the acetylation reactions of the biocatalysts using a single-stage batch system. After each reaction (8 h reaction time: immobilized LIPB and Novozym 435 and 12 h reaction time: lyophilized LIPB, solution commercial CALB and immobilized CALB), the experiment was halted and the reaction medium was centrifuged. The liquid phase was removed and analyzed by HPLC to determine the conversion rate and enantiomeric excess, while the biocatalyst was washed with ethyl acetate to remove any product and substrate from the previous reaction. Then, the lipase was put through the next reaction. In each run a 100 μL sample was taken to analyze remaining hydrolytic activity. The biocatalyst was reused 12-fold in a batch system.

3. Results and discussion

3.1. Production of lipase B from C. antarctica (LIPB) in P. pastoris

After the improvements reported in the work of Maurer and collaborators,⁵⁰ the condition of 100 g L⁻¹ of glycerol as carbon source was chosen as the best condition for this study. Fig. 2 represents the growth of the yeast Pichia pastoris in a conventional batch system at 30 °C, pH 7.0, 30% oxygen saturation, and 100 g L⁻¹ of glycerol.

Fig. 2 Kinect of biomass growth, lipase production using p-NFL as substrate and consumption of glycerol and nitrogen on a batch with minimum media using 100 g L⁻¹ of glycerol at 30 °C, pH 7.0 and cascade varying aeration (0–3 L min⁻¹) and stirrer (250–700 rotation per minute – rpm) to control DO at 30%.

It is noteworthy that the production of the lipase is associated with the growth of the biomass as expected since the lipase gene is combined with a constitutive promoter. The lag phase ends after 18 h and the glycerol ends at 42 h. The nitrogen is kept stable during the whole process, indicating that this is not a limiting growth factor.

In Table 2, the main parameters used to evaluate the efficiency of the batch and production at the end of fermentation are shown.

Table 2 Kinetic parameters of the growth of Pichia pastoris yeast transformed with pPGKΔ3_PRO_LIPB and main parameters of lipase production using tributyrin as substrate at 42 h

Fermentation time (h)	42
Yx/s	0.44 ± 0.09
Yp/s	314 ± 51
Qp (U g^−1 h^−1)	7.15 ± 1.17
Activity (U L^−1)	13 903 ± 609
Specific activity (U mg^−1)	55 ± 3

---

Lipase production utilizing the aforementioned promoter is not as high as the one achieved employing the most strong promoter usually employed on papers (pAOX), it is 3.5 fold lower when compared with the work made by Vadhana and collaborators.⁶⁴ However, the advantage of using a constitutive promoter (PGK) rather than an inducible promoter (pAOX) may compensate this lower lipase production. It must be considered that in large scale the storing and the use of methanol may promote some difficulties.⁶⁵

Candida antarctica lipase B was produced successfully in Pichia pastoris using a PGK constitutive promoter, but better production strategies (fed-batch and continuous) and genetic engineering (heterologous constructions with multiple copies of the synthetic gene) are being studied aiming to increase the productivity of the process to make it competitive compared to inductive expression.

After 42 h of growth, the cells were removed by centrifugation and the supernatant containing the enzyme (13 903 U_t L⁻¹, volumetric activity) was lyophilized for further immobilization experiments.

3.2. Studies into the immobilization of CALB and LIPB on core–shell supports and Accurel MP 1000

The loading capacity of the different supports using LIPB and CALB was studied (see on ESI: Table 1S and Fig. 1S to 8S†). Under low enzymatic loading, almost all protein (and activity) becomes adsorbed on all the supports (more than 90–95%). When the amount of protein increased, and around 60 mg of enzyme are added to one gram of support, the obtained results varied, depending on the enzyme and the support. Using Accurel, 75% of the enzyme activity were adsorbed after 24 h, while only 40% of the protein were adsorbed, using both LIPB and CALB, indicating that some purification of the enzyme is being achieved. Comparing PS-co-DVB/PS-co-DVB-LIPB and Accurel-LIPB, the immobilization process is slightly higher both in protein and activity basis, while using CALB the enzymatic activity decrease to 60% and the protein remains similar. Comparing PMMA-co-DVB/PMMA-co-DVB-LIPB and PS-co-DVB/PS-co-DVB-LIPB the obtained results are quite similar using LIPB while using CALB the immobilization yield dropped to less of 45%, with a lower decrease in protein adsorption. Analyzing the obtained results for CALB in protein bases, it is observed that employing an excess of enzyme the amount of immobilized enzyme is similar using Accurel and PS-co-DVB/PS-co-DVB, but not for PMMA-co-DVB/PMMA-co-DVB that have significantly lower specific area. It is important to point that for all four supports, the pores are wide enough for the adsorption of small CALB onto the internal part of the shell. However, the contaminants may be expected to be different for both lipases extracts (and that may justify the different behavior when overloading the support, even when the target protein should behave in a very similar way). Considering CALB solution, a large contaminant protein, which is absent on LIPB preparation, could close the pores of the particles.²² Thus, the loaded preparations of LIPB presented some more enzyme than that of CALB, maximizing the difference using, for example, PMMA-co-DVB/PMMA-co-DVB (500 units per g using LIPB but 300 using CALB) and minimizing the differences using Accurel (525 units per g using LIPB, 514 using CALB). We decided to compare all the preparations to ensure that the differences between the enzymes are due to the enzyme and not to some artifact caused by a better immobilization of one enzyme in one specific support, taking advantage of the differences in the surface morphology and hydrophobicity of the 4 supports.⁶⁶

3.3. Esterification activity of the core–shell derivatives

The biocatalysts produced in this study were evaluated as catalysts in oleic acid and ethanol esterification reactions in solvent-free medium. Table 3 shows the results comparing the 4 biocatalysts of LIPB, the ones from CALB and the commercial Novozym 435, a biocatalyst recognized as one of the most efficient ones. PS-co-DVB/PS-co-DVB-LIPB has the highest activity, similar to Novozym 435. Accurel and PMMA-co-DVB/PMMA-co-DVB produced LIPB biocatalyst quite less active in this reaction, even though both of them have similar enzyme loading. PMMA/PMMA-LIPB was the least active one. CALB immobilized on all supports was less active than the corresponding counterparts of LIPB. Differences were maximal using PS-co-DVB/PS-co-DVB (almost 3 folds) and minimal using PMMA/PMMA (just a 5%), suggesting strong modulation of the differences between the two enzymes with the support. The most active preparation using CALB was Novozym 435, but even this preparation was less active than PS-co-DVB/PS-co-DVB-LIPB (just by a 8%).

Table 3 Esterification activity of LIPB and CALB immobilized on different polymeric supports

Supports	Esterification activity (U gsupport^−1)
LIPB
Accurel MP 1000	1.97 000 ± 220
PS-co-DVB/PS-co-DVB	2.95 800 ± 109
PMMA-co-DVB/PMMA-co-DVB	1.75 500 ± 170
PMMA/PMMA	1.35 500 ± 170
Novozyme 435	2.72 800 ± 339
CALB
Accurel MP 1000	1.77 000 ± 150
PS-co-DVB/PS-co-DVB	1.01 800 ± 109
PMMA-co-DVB/PMMA-co-DVB	1.81 500 ± 105
PMMA/PMMA	1.29 800 ± 105
Novozyme 435	2.72 800 ± 339

---

3.4. Kinetic resolution of DL-1 using recombinant LIPB immobilized on different supports

The kinetic resolution of DL-1 (Fig. 3) was studied using batch systems employing both enzymes in different forms: lyophilized, immobilized on Accurel and on the core–shell supports and Novozyme 435, used as reference (a biocatalyst that has been previously successfully employed in this reaction).^63,67 The reaction conditions were determined based on the best condition found in the optimization of the reaction using Novozym 435,⁵⁵ which were 45 °C, 4 mg mL⁻¹ of substrate, and 222.8 U of enzyme, with hexane as a solvent. The reactions were halted when the yield reached 50% (8 h (see ref. 67) or 12 h, depending on the catalyst). Table 4 resumes main results.

Fig. 3 Schematic diagram of the enantioselective reaction of DL-1 catalyzed by lipase biocatalyst. The reaction condition was: 8 mg substrate dissolved in 2.0 mL vinyl acetate in hexane (1 : 5.5), 222.8 U of biocatalyst at 45 °C, with 8 h reaction time for the core–shell derivatives and Novozyme 435, and 8 h of reaction time (immobilized LIPB and Novozym 435) or 12 h (lyophilized LIPB, solution commercial CALB and immobilized CALB).

Table 4 Results of racemic resolution of DL-1 with CALB and LIPB immobilized on different supports

Commercial lipase	X (%)	eep	E	Initial velocity (μmol min^−1 g^−1) × 10^2
a Batch conditions: 45 °C; 4 mg mL^−1 of substrate; 222.8 U of enzyme; 8 h. Hexane as solvent.b 12 h reaction time.
Lyophilized CALB	20^b ± 0.1	99	>200	3.30
CALB immobilized on Accurel MP 1000	40.6^b ± 0.2	99	>200	6.71
CALB immobilized on PS-co-DVB/PS-co-DVB	42.0^b ± 0.1	99	>200	6.94
CALB immobilized on PMMA-co-DVB/PMMA-co-DVB	40.0^b ± 0.2	99	>200	6.61
CALB immobilized on PMMA/PMMA	31.0^b ± 0.5	99	>200	3.21
Novozyme 435	49.7^a ± 0.1	99	>200	8.21

---

Recombinant lipase	X (%)	eep	E	Initial velocity (μmol min^−1 g^−1) × 10^2
Lyophilized LIPB	3.0^b ± 0.5	99	>200	0.74
LIPB immobilized on Accurel MP 1000	41.0^a ± 0.5	99	>200	10.16
LIPB immobilized on PS-co-DVB/PS-co-DVB	47.5^a ± 0.1	99	>200	11.77
LIPB immobilized on PMMA-co-DVB/PMMA-co-DVB	41.2^a ± 0.1	99	>200	10.21
LIPB immobilized on PMMA/PMMA	36.2^a ± 0.1	99	>200	5.30

---

All catalysts were capable of catalyzing the acetylation of DL-1 to obtain the product L-(−)-2 with high ee_p (99%), and a high enantiospecificity (E > 200). The main difference was in the initial activity.

Analyzing the group of the produced biocatalysts in which CALB was employed, the highest activity is obtained using the commercial biocatalyst. The lyophilized enzyme is the least active form of the lipase, as may be expected because it will be forming aggregates, except PMMA/PMMA-CALB that was significantly less active the other home-made preparations. All the other preparations offered very close activity, even though the enzyme loading was higher for Accurel and minimal for PS-co-DVB/PS-co-DVB.

The lyophilized LIPB is even less active than CALB, however this behavior is a result of many factors, as the commercial preparation may have different components compared with the home-made lipase. Except PMMA/PMMA-LIPB, all immobilized preparations were found to be slightly more active (25%) than the commercial Novozym 435. All the home-made preparations counterparts prepared using LIPB were more active than using CALB, in some cases by more than 50% (e.g., 65% using PMMA/PMMA). Using Accurel or PS-co-DVB/PS-co-DVB, where the enzyme loadings are similar using both enzymes, the improvement in activity of this recombinant enzyme compared to the commercial one is 50%. Curiously, using PMMA-co-DVB/PMMA-co-DVB, the differences in activity almost matched the differences in loading.

3.5. Kinetic resolution of DL-2 using recombinant LIPB and CALB immobilized on different supports

Next, the kinetic resolution of DL-2 (Fig. 4) was analyzed using batch systems with the same formulations of Section 3.4. The reaction conditions were the ones optimized during the experimental design for DL-2 and Novozym 435,⁶³ which were 45 °C, 5 mg mL⁻¹ of substrate (using vinyl acetate as acylating agent, solvent free medium), 26 mg of biocatalyst. Table 5 shows the most relevant results.

Fig. 4 Schematic diagram of the enantioselective reaction of DL-2 catalyzed by commercial CALB and home-made LIPB immobilized on different supports. The condition was: 10 mg substrate dissolved in 2.0 mL vinyl acetate (solvent-free medium) at 45 °C and 71 U of biocatalyst.

Table 5 Results of the enantioselective resolution of DL-2 for the different versions of Candida antarctica lipase B

CALB (commercial)	X (%)	eep	E	Initial velocity (μmol min^−1 g^−1) × 10^2
a Batch conditions: 45 °C; 5 mg mL^−1 of substrate; 71 U of enzyme; 8 h reaction time; solvent-free reaction; vinyl acetate as acylating agent and solvent.b Batch conditions: 45 °C; 5 mg mL^−1 of substrate; 71 U of enzyme; 13 h reaction time; solvent-free reaction; vinyl acetate as acylating agent and solvent.
Lyophilized CALB	28.0^b ± 0.8	99	>200	4.93
CALB immobilized on Accurel MP 1000	40.8^b ± 0.1	99	>200	7.18
CALB immobilized on PS-co-DVB/PS-co-DVB	45.3^a ± 0.5	99	>200	7.97
CALB immobilized on PMMA-co-DVB/PMMA-co-DVB	43.3^a ± 0.2	99	>200	7.62
CALB immobilized on PMMA/PMMA	33.0^b ± 0.5	99	>200	4.21
Novozyme 435	49.0^a ± 0.1	99	>200	8.62

---

Recombinant lipase	X (%)	eep	E	Initial velocity (μmol min^−1 g^−1) × 10^2
Lyophilized LIPB	18.0^b ± 0.8	99	>200	4.75
LIPB immobilized on Accurel MP 1000	48.0^a ± 0.5	99	>200	12.67
LIPB immobilized on PS-co-DVB/PS-co-DVB	49.3^a ± 0.5	99	>200	13.01
LIPB immobilized on PMMA-co-DVB/PMMA-co-DVB	48.6^a ± 0.2	99	>200	12.83
LIPB immobilized on PMMA/PMMA	36.0^b ± 0.5	99	>200	5.88

---

Again, E and ee_p were very high for all catalysts but there were clear differences in activity.

The lyophilized preparations of both enzymes have similar activities and are the lowest ones, as mentioned before, because the use of an aggregated enzyme may offer complex results and difficult the understanding, although it is possible to assume the existence of diffusional problems. CALB offered the highest activity using the commercial preparation, shortly followed by the home-made preparations, except for PMMA/PMMA-CALB that was less than 50% active. Again, the activity of CALB immobilized on PMMA-co-DVB/PMMA-co-DVB was higher than should be expected from the low loading of the enzyme in the support, almost matching the activity of the other home-made preparations when the loading was almost 60% of that of the other preparations.

The use of the biocatalysts from LIPB provided higher activities than using CALB, the difference in activity is more than 50% for all preparations, and the activity comparing with Novozym 435 is more than 50% in the case of PS-co-DVB/PS-co-DVB-LIPB. Again in the case of PMMA-co-DVB/PMMA-co-DVB, the difference in activity almost matches the difference in loading, suggesting a similar specific activity of the enzyme when immobilized on these support.

The immobilized LIPB proved more efficient in the enantioselective resolution of DL-2 than the derivatives from commercial CALB, including Novozyme 435, since it achieved maximum conversion in a shorter time (8 h). The higher conversion of DL-2 than of DL-1 could be due to the fact that the structure of DL-2 is less sterically hindered than the structure of DL-1 (in this case, the presence of benzyl groups could have hampered the interaction between the substrate and the biocatalyst).

The product, L-5, obtained from Novozyme 435 and LIPB in PS-co-DVB (core and shell) was found to have very high enantiomeric excess (ee_p of 99%) and high enantiomeric ratio (E).

These data gave an improvement of a 50% of the productivity (considering the whole reaction cycle) using PS-co-DVB-LIPB instead of Novozym 435 using DL-2 as substrate (0.060 mg_product g_derivative⁻¹ h⁻¹).

3.6. Reuse of the biocatalyst

Reuse studies were conducted on the recombinant LIPB immobilized on Accurel MP 1000, PS-co-DVB/PS-co-DVB, PMMA-co-DVB/PMMA-co-DVB and PMMA–PMMA and compared with commercial biocatalyst Novozyme 435, for the kinetic resolution of DL-2. The reactions were run for 4 days under the following optimized conditions: 5 mg mL⁻¹ of substrate and 26 mg derivative, 8 h reaction time, solvent-free reaction, vinyl acetate was simultaneously acylating agent and solvent.

After 12 reaction cycles (considering an 8 h reaction) no significant loss in the conversion of DL-2 was found when the recombinant LIPB derivative immobilized on PS-co-DVB was used throughout all the cycles conducted. The other derivatives suffered a slight decrease in conversion after the 7th reaction cycle, including Novozym 435 (Fig. 5).

Fig. 5 Operational stability in the kinetic resolution of DL-2 using different derivatives under the following conditions: 5 mg mL⁻¹ of substrate; 71 U of enzyme; at 45 °C; reaction time of 8 hours in a batch system. LIPB on PS-co-DVB (core and shell)-square; LIPB on PMMA-co-DVB (core and shell)-circle; LIPB on Accurel MP 1000-triangle, LIPB on PMMA (core and shell)-star and Novozym 435-diamond.

All preparations could be stored at 4–30 °C for 60 days and keeping more than 95% of the initial activity.

4. Conclusion

The recombinant lipase B from Candida antarctica (LIPB) expressed in Pichia pastoris using the constitutive promoter PGK has proved to be a very useful alternative to the commercial CALB in certain cases. The production of ethyl oleate is slightly lower for the new enzyme. However, in the resolution of DL-1 and DL-2 via transesterification (using different media), LIPB immobilized in Accurel or PS-co-DVB/PS-co-DVB presented more activity per enzyme molecule than CALB immobilized in similar supports, while when immobilized in PMMA-co-DVB/PMMA-co-DVB the activity of both enzymes were similar.

The recombinant LIPB immobilized on PS-co-DVB proved to be the most efficient in the enantioselective resolution of both racemic derivatives, DL-1 and DL-2. The productivity for DL-2 resolution was 50% higher than the commercial Novozym 435, and the new derivative was operationally more stable than Novozyme 435. The products obtained had a high level of purity (ee of 99% for both derivatives). Both products of the enantioselective reaction, L-2 and L-5, obtained from the racemic derivatives (DL-1 and DL-2, respectively), are intermediates from different pharmacological pathways involved in the synthesis of building blocks for drugs that inhibit the etiological agent of Chagas disease, Trypanosoma cruzi. These results constitute a successful, novel approach for simultaneously evaluating new technologies for producing, immobilizing and using Candida antarctica lipase B.

The better enzyme performance on these reactions has encouraged our group to analyze the differences between LIPB and CALB. At first glance, the expression of the enzyme in Pichia may gave a more glycosylated protein, but preliminary results suggest a less compact structure of the new enzyme; the full characterization of this enzyme is under study and will be the subject of a forthcoming paper.

• The results shown in this paper are protected by the Brazilian patent application number US 20110183400 A1.⁴⁹

Acknowledgements

The authors thank CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico), CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) and FAPERJ (Fundação Carlos Chagas Filho de Apoio à Pesquisa do Estado do Rio de Janeiro) for the scholarships and the financial support. R. Fernandez-Lafuente thanks MINECO of Spanish Government, by the grant CTQ2013-41507-R and CNPq (PVE 301139/2014-8). The help and suggestions during the writing of the paper by Dr Ángel Berenguer are gratefully recognized (Universidad de Alicante).

References

1. E. M. Anderson, K. M. Larsson and O. Kirk, Biocatal. Biotransform., 1998, 16, 181.
2. V. Gotor-Fernández, E. Busto and V. Gotor, Adv. Synth. Catal., 2006, 348, 797.
3. U. T. Bornscheuer and R. J. Kazlauskas, Hydrolases in Organic Synthesis: Regio- and Stereoselective Biotransformations, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2nd edn, 2006.
4. Z. Boros, P. Falus, M. Márkus, D. Weiser, M. Oláh, G. Hornyánszky, J. E. Nagy and L. Poppea, J. Mol. Catal. B: Enzym., 2013, 85–86, 119.
5. A. G. Cunha, M. D. Besteti, E. A. Manoel, A. A. T. Silva, R. V. Almeida, A. B. C. Simas, R. Fernandez-Lafuente, J. C. Pinto and D. M. G. Freire, J. Mol. Catal. B: Enzym., 2014, 100, 59.
6. P. J. Dunn, A. S. Wells and M. T. Williams, Green Chemistry in the Pharmaceutical Industry, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2010.
7. D. Goswami, J. K. Basu and S. De, Crit. Rev. Biotechnol., 2013, 33, 81.
8. M. Kapoor and M. N. Gupta, Process Biochem., 2012, 47, 555.
9. O. Kirk and M. W. Christensen, Org. Process Res. Dev., 2002, 6, 446.
10. D. M. Solano, P. Hoyos, M. J. Hernaiz, A. R. Alcantara and J. M. Sanchez-Montero, Bioresour. Technol., 2012, 115, 196.
11. M. M. Soumanou, M. Pérignon and P. Villeneuve, Eur. J. Lipid Sci. Technol., 2013, 115, 270.
12. W. Du, W. Xu, D. Liu and J. Zeng, J. Mol. Catal. B: Enzym., 2004, 30, 125.
13. A. Khan, S. K. Sharma, A. Kumar, A. C. Watterson, J. Kumar and V. S. Parmar, ChemSusChem, 2014, 7, 379.
14. G. Potvin, A. Ahmad and Z. Zhang, Biochem. Eng. J., 2012, 64, 91.
15. S. Macauley-Patrick, M. L. Fazenda, B. McNeil and L. M. Harvey, Yeast, 2005, 22, 249.
16. A. L. Zhang, J. X. Luo, T. Y. Zhang, Y. W. Pan, Y. W. H. Tan, C. Y. Fu and F. Z. Tu, Mol. Biol. Rep., 2009, 36, 1611.
17. X. Qin, J. Qian, G. Yao, Y. Zhuang, S. Zhang and J. Chu, Appl. Environ. Microbiol., 2011, 77, 3600.
18. G. Stadlmayr, A. Mecklenbräuker, M. Rothmüller, M. Maurer, M. Sauer, D. Mattanovich and B. Gasser, J. Biotechnol., 2010, 150, 519.
19. D. L. Gazzoni, Glicerina, biocombustíveis e bioprodutos, http://www.gazzoni.eng.br/pagina42.htm, 02 April 2015.
20. P. Adlercreutz, Chem. Soc. Rev., 2013, 42, 6406.
21. R. Dicosimo, J. McAuliffe, A. J. Poulose and G. Bohlmann, Chem. Soc. Rev., 2013, 42, 6437.
22. C. Garcia-Galan, A. Berenguer-Murcia, R. Fernandez-Lafuente and R. C. Rodrigues, Adv. Synth. Catal., 2011, 353, 2885.
23. U. Guzik, K. Hupert-Kocurek and D. Wojcieszynska, Molecules, 2014, 19, 8995.
24. E. T. Hwang and M. B. Gu, Eng. Life Sci., 2013, 13, 49; K. Min and Y. J. Yoo, Biotechnol. Bioprocess Eng., 2014, 19, 553–567.
25. R. C. Rodrigues, C. Ortiz, A. Berenguer-Murcia, R. Torres and R. Fernández-Lafuente, Chem. Soc. Rev., 2013, 42, 62907.
26. R. A. Sheldon and S. van Pelt, Chem. Soc. Rev., 2013, 42, 6223.
27. O. Barbosa, C. Ortiz, A. Berenguer-Murcia, R. Torres, R. C. Rodrigues and R. Fernandez-Lafuente, Biotechnol. Adv., 2015, 33, 435.
28. R. Fernandez-Lafuente, P. Armisén, P. Sabuquillo, G. Fernández-Lorente and J. M. Guisán, Chem. Phys. Lipids, 1998, 93, 185.
29. J. M. Palomo, G. Muoz, G. Fernández-Lorente, C. Mateo, R. Fernández-Lafuente and J. M. Guisán, J. Mol. Catal. B: Enzym., 2002, 19–20, 279.
30. E. A. Manoel, M. F. P. Ribeiro, J. C. S. Dos Santos, M. A. Z. Coelho, A. B. C. Simas, R. Fernandez-Lafuente and D. M. G. Freire, Process Biochem., 2015, 50, 1557.
31. A. M. Brzozowski, U. Derewenda, Z. S. Derewenda, G. G. Dodson, D. M. Lawson and J. P. Turkenburg, et al., Nature, 1991, 351, 491.
32. K. K. Kim, H. K. Song, D. H. Shin, K. Y. Hwang and S. W. Suh, Structure, 1997, 5, 173.
33. K. E. Jaeger, S. Ransac, H. B. Koch, F. Ferrato and B. W. Dijkstra, FEBS Lett., 1993, 332, 143.
34. M. Cygler and J. D. Schrag, Biochim. Biophys. Acta, 1999, 1441, 205.
35. R. Verger, Trends Biotechnol., 1997, 15, 32.
36. P. Reis, K. Holmberg, H. Watzke and M. E. R. Miller, Adv. Colloid Interface Sci., 2009, 147–148, 237–250.
37. R. Fernandez-Lafuente, P. Armisén, P. Sabuquillo, G. Fernández-Lorente and M. Guisán, Chem. Phys. Lipids, 1998, 93, 185.
38. J. M. Palomo, C. Ortiz, G. Fernández-Lorente, M. Fuentes, J. M. Guisán and R. Fernández-Lafuente, Enzyme Microb. Technol., 2005, 36, 447.
39. G. Fernández-Lorente, J. M. Palomo, M. Fuentes, C. Mateo, J. M. Guisán and R. Fernández-Lafuente, Biotechnol. Bioeng., 2003, 82, 232.
40. J. M. Palomo, M. M. Peñas, G. Fernández-Lorente, C. Mateo, A. G. Pisabarro, L. Ramírez, R. Fernández-Lafuente and J. M. Guisán, Biomacromolecules, 2003, 4, 204.
41. R. V. Almeida, R. V. Branco, R. V. B. Peixoto, C. Lima, S. M. C. Martins, O. B. O. Antunes and D. M. G. Freire, Biochem. Eng. J., 2008, 39, 531.
42. J. V. Bevilaqua, A. G. Cunha, L. M. Lima, T. L. M. Alves, E. J. Barreiro, L. M. C. Paiva and D. M. G. Freire, Appl. Biochem. Biotechnol., 2005, 121, 117.
43. A. G. Cunha, G. Fernández-Lorente, M. L. E. Gutarra, J. V. Bevilaqua, R. V. Almeida, L. M. C. Paiva, R. Fernández-Lafuente, J. M. Guisán and D. M. G. Freire, Appl. Biochem. Biotechnol., 2009, 156, 563.
44. E. A. Manoel, J. C. S. dos Santos, D. M. G. Freire, N. Rueda and R. Fernandez-Lafuente, Enzyme Microb. Technol., 2015, 71, 53.
45. M. D. Besteti, D. M. G. Freire and J. C. Pinto, Macromol. React. Eng., 2011, 5, 518.
46. M. D. Besteti, A. G. Cunha, D. M. G. Freire and J. C. Pinto, Macromol. Mater. Eng., 2014, 299, 135.
47. M. K. Lenzi, F. M. Silva, E. L. Lima and J. C. Pinto, J. Appl. Polym. Sci., 2003, 89, 3021.
48. M. C. C. Pinto, D. M. G. Freire and J. C. Pinto, Molecules, 2014, 19, 12509.
49. A. M. D. Castro, J. V. Bevilaqua, D. M. G. Freire, F. A. Torres, L. M. M. Santanna, M. L. E. Gutarra, C. A. Barbosa, R. V. Almeida, R. R. de Menezes and A. G. Cunha, A1-process for production of lipases by genetic modification of yeast, US 20110183400, 2011.
50. M. Maurer, M. Kühleitner, B. Gasser and D. Mattanovich, Microb. Cell Fact., 2006, 5, 37.
51. J. K. Fawcett and J. E. Scott, J. Clin. Pathol., 1960, 13, 156.
52. A. Tabacco, F. Meiattini, E. Moda and P. Tarli, Clin. Chem., 1979, 25, 336.
53. M. Bradford, Anal. Biochem., 1976, 72, 248.
54. C. Garcia-Galan, O. Barbosa, K. Hernandez, J. C. S. Dos Santos, R. C. Rodrigues and R. Fernandez-Lafuente, Molecules, 2014, 19, 7629.
55. K. Hernandez, C. Garcia-Galan and R. Fernandez-Lafuente, Enzyme Microb. Technol., 2011, 49, 72.
56. R. C. Rodrigues, K. Hernandez, O. Barbosa, N. Rueda, C. Garcia-Galan, J. C. S. Dos Santos, A. Berenguer-Murcia and R. Fernandez-Lafuente, Curr. Org. Chem., 2015, 19, 1707.
57. K. Yamabe, T. Ihara and A. Jyo, Sep. Sci. Technol., 2001, 3511–3528.
58. J. Wang, C. Chen, S. M. Buck and Z. Chen, J. Phys. Chem. B, 2001, 12118–12125.
59. N. Gupta, P. Rathi, R. Singh, V. K. Goswami and R. Gupta, Appl. Microbiol. Biotechnol., 2005, 67, 648.
60. R. Scherer, J. V. Oliveira, S. Pergher and D. de Oliveira, Mater. Res., 2011, 14, 483.
61. E. Séverac, O. Galy, F. Turon, C. A. Pantel, J. S. Condoret, P. Monsan and A. Marty, Enzyme Microb. Technol., 2011, 48, 61.
62. E. A. Manoel, K. C. Pais, A. G. Cunha, A. B. C. Simas, M. A. Z. Coelho and D. M. G. Freire, Org. Process Res. Dev., 2012, 16, 1378.
63. A. G. Cunha, A. A. T. Da Silva, M. G. Godoy, R. V. Almeida, A. B. C. Simas and D. M. G. Freire, J. Chem. Technol. Biotechnol., 2012, 88, 205.
64. A. K. P. Vadhana, P. Samuel, R. M. Berin, J. Krishna, K. Kamatchi and S. Meenakshisundaram, Enzyme Microb. Technol., 2013, 52, 177.
65. J. C. Goodrick, M. Xu, R. Finnegan, B. M. Schilling, S. Schiavi, H. Hoppe and N. C. Wan, Biotechnol. Bioeng., 2001, 74, 492.
66. G. Fernandez-Lorente, Z. Cabrera, G. Godoy, R. Fernandez-Lafuente, J. M. Palomo and J. M. Guisan, Process Biochem., 2008, 43, 1061.
67. E. A. Manoel, K. C. Pais, A. G. Cunha, M. A. Z. Coelho, D. M. G. Freire and A. B. C. Simas, Tetrahedron: Asymmetry, 2012, 23, 47.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra22508f

---