Design of a core–shell support to improve lipase features by immobilization

Abstract

Different core–shell polymeric supports, exhibiting different morphologies and composition, were produced through simultaneous suspension and emulsion polymerization, using styrene (S) and divinylbenzene (DVB) as co-monomers. Supports composed of polystyrene in both the core and the shell (PS/PS) and the new poly(styrene-co-divinylbenzene) support (PS-co-DVB/PS-co-DVB) were used for the immobilization of three different lipases (from Rhizomucor miehie (RML), from Themomyces lanuginosus (TLL) and the form B from Candida antarctica, (CALB)) and of the phospholipase Lecitase Ultra (LU). The features of the new biocatalysts were evaluated and compared to the properties of commercial biocatalysts (Novozym 435 (CALB), Lipozyme RM IM and Lipozyme TL IM) and biocatalysts prepared by enzyme immobilization onto commercial octyl-agarose, a support reported as very suitable for lipase immobilization. It was shown that protein loading and stability of the biocatalysts prepared with the core–shell supports were higher than the ones obtained with commercial octyl-agarose or the commercial lipase preparations. Besides, it was shown that the biocatalysts prepared with the core–shell supports also presented higher activities than commercial biocatalysts when employing different substrates, encouraging the use of the produced core–shell supports for immobilization of lipases and the development of new applications.

---

1. Introduction

Enzyme immobilization enables the recovery and the reuse of these expensive biocatalysts as long as the preparation is stable enough.^1–4 Therefore, in order to benefit from this challenge, many efforts have been devoted to turn immobilization into the solution for other enzyme limitations, such as stability, activity, selectivity, specificity or purity.^5–12

Lipases are among the most used enzymes in biocatalysis, due to their characteristic wide specificity and the wide range of reactions that these enzymes can catalyze (including many promiscuous reactions).^13–15 Besides, lipases show very high enantioselectivity^16–19 and are very robust, being successfully employed in different reaction media (e.g., aqueous media, organic solvents, neoteric media).^20,21

The active centers of most lipases are secluded from the reaction media by a polypeptide chain (the lid), which is isolated from the medium by the large hydrophobic pocket where the active center is located (closed form).^23–27 The lid can move and exposes this hydrophobic pocket to the medium, generating the open and active form of the lipase. This open lipase form readily adsorbs onto hydrophobic surfaces, including oil drops,^22,28 hydrophobic supports,²⁹ other open lipase molecules,^30,31 other hydrophobic proteins.³²

In fact, a strategy that is becoming very popular for the immobilization of lipases is based on the interfacial activation of the enzyme on hydrophobic support surfaces.³³ This strategy allows the immobilization, purification and stabilization of the open lipase form (usually leading to hyper-activation), also producing an increase in the lipase stability for this reason.²⁹ It has been reported that the internal morphology or physical properties of the hydrophobic support may tune the lipase properties, including its activity, stability and specificity.^34–36 Therefore, there is a great interest in the new development of new hydrophobic matrices that can further improve lipase properties. Among these new materials, core–shell polymeric particles, consisting of “large” particles (core) coated with small nano-particles (shell) may play a pivotal role.^37–40

Different techniques have been employed for the production of core–shell particles. Among them, the combined suspension and emulsion polymerization process has special interest. This technique typically comprises two fundamental steps. In the first step (suspension polymerization), the particle cores are synthesized. When the monomer conversion reaches a certain value in the core formation, the second step is initialized. To this goal, the elements of a typical emulsion polymerization are fed into the reaction vessel. The new nanoparticles coagulate over the previously prepared particle cores to form the shell. During the second reaction step, the suspension and emulsion polymerization processes are conducted simultaneously. At the end of the process, micrometric, porous and (in some cases) functionalized polymer particles are obtained.^40,41

In the present manuscript, distinct polymeric supports presenting core–shell morphology were produced through simultaneous suspension and emulsion polymerization, using styrene (S) and divinylbenzene (DVB) as co-monomers. S and DVB are hydrophobic monomers, while DVB can also promote chain crosslinking, leading to modification of the morphology and mechanical resistance of the obtained polymer particles. The first step of this study comprised the determination of how the co-monomers feed flow rate of S and DVB affects the specific area and porosity of the synthesized core–shell particles. Afterwards, among the different supports that were produced, one of them was selected for the enzyme immobilization procedure: the support with the highest specific area (PS-co-DVB/PS-co-DVB). The previously described core–shell polystyrene support (PS/PS), that has been successfully employed in the immobilization of the lipase B from Candida antarctica, was also employed for comparison.⁴¹

The prepared polymeric supports were used for the immobilization of three lipases: lipases from Rhizomucor miehei, RML,⁴² from Thermomyces lanuginosus, TLL,⁴³ and the form B from Candida antarctica, CALB.⁴⁴ While CALB has a very small lid, which does not completely isolate its active center from the reaction medium,⁴⁵ TLL and RML exhibit very large lids.^46,47 The polymeric supports were also used for immobilization of the chimeric artificial phospholipase Lecitase Ultra (LU).^48,49 PS/PS core–shell supports had been previously and successfully used for CALB immobilization, but this is the first attempt to immobilize the other enzymes on this support. The high specific area PS-co-DVB/PS-co-DVB support is used to immobilize enzymes for the first time in this paper.

The new biocatalysts were compared with commercial biocatalysts (Novozym 435 (CALB), Lipozyme RM IM and Lipozyme TL IM) and biocatalysts prepared through enzyme immobilization onto commercial octyl-agarose, a very popular support successfully used in many instances for lipase immobilization.^33,50 Lipases immobilized on octyl-agarose have been reported to be much more stable than the free enzyme, and even more than some lipases immobilized via multipoint covalent attachment.^51,52 This has been explained by the higher stability of the adsorbed open form of the lipase compared to the lipase in conformational equilibrium.^53,54 Lipases tend to form dimeric aggregates with altered properties and that may alter the results of the activity and stability studies and suggests that the use of free lipase to compare the properties with immobilized enzymes may not be very adequate.^55–59

2. Materials and methods

2.1. Materials

Solutions of CALB (19.11 mg of protein per mL), TLL (36 mg of protein per mL), RML (13.7 mg of protein per mL) and of LU (16 mg of protein per mL), and the commercial immobilized biocatalysts Novozym 435®, Lipozyme® TL IM and Lipozyme® RM IM were kindly provided by Novozymes (Spain). Octyl-sepharose (octyl-agarose) beads were purchased from GE Healthcare. Methyl mandelate, p-nitrophenyl butyrate (p-NPB) and triacetin were obtained from Sigma Chemical Co. (USA).

Styrene supplied by Sigma Aldrich (USA) (minimum purity of 99.5% (wt/wt)) was used as monomer for the production of PS/PS particles. For the production of PS-co-DVB/PS-co-DVB particles, styrene was provided by INOVA (Brazil) and distilled under vacuum before its use and DVB was supplied by Merck (USA).

Other reagents and solvents were of analytical grade and were used as received, without any purification step.

2.2. Preparation of core–shell polymeric supports and their characterization

Core–shell polymeric supports were synthesized through the combined suspension–emulsion polymerization process.^{40,41,60–62} The procedures used for production of reference support PS/PS particles have been presented elsewhere.⁶⁰ However, modifications of the original procedures were proposed here in order to increase the specific area and porosity of the obtained polymeric particles.

Reactions were carried out in an open 1 L jacketed glass reactor (FGG Equipamentos Científicos, São Paulo, Brazil) equipped with a thermostatic bath (Haake Phoenix II model, Thermo Scientific, Karlsruhe, Germany) that was employed to keep the reactor temperature at 85 °C. For the production of PS/PS particles, styrene was used as the only monomer in the suspension and emulsion processes.⁶⁰ For the production of PS-co-DVB/PS-co-DVB, copolymerization of styrene (75% (wt/wt) and DVB (25% wt/wt)) was conducted during the suspension and emulsion polymerization steps. For production of PS-co-DVB individual core particles, classic suspension copolymerization was performed, without addition of the emulsion feed.

Initially, 100 g of an organic solution (containing the monomer mixture and 3.8% (wt/wt) of the initiator benzoyl peroxide) were dispersed in 370 g of an aqueous solution (containing distilled water and 0.80% (wt/wt) of poly(vinyl alcohol), used as stabilizer). The dispersion was kept under continuous agitation (950 rpm in the PS/PS reaction and 800 rpm in PS-co-DVB and PS-co-DVB/PS-co-DVB polymerizations) at a constant temperature of 85 °C. After two hours of reaction, the emulsion constituents (the monomer mixture and the aqueous solution, containing distilled water, 0.13% (wt/wt) of the initiator potassium persulfate, 0.13% (wt/wt)% of sodium bicarbonate and 1% (wt/wt) of sodium lauryl sulfate) were added to the reaction medium. 30 g of the monomer mixture and 230 g of the aqueous solution were added as a single load. The remaining 70 g of the monomer mixture were added under continuous flow (0.026 L h⁻¹) in the PS/PS reaction.⁶⁰ For the production of PS-co-DVB/PS-co-DVB particles, different flow rates were employed, as shown in Table 1. After feeding, two additional hours of reaction were permitted to ensure the appropriate coverage of the core and formation of the shell. Then, the reactor was cooled down to room temperature and the obtained particles were filtrated and washed with cold water. Finally, the obtained polymer particles were dried in a vacuum oven at 30 °C until constant mass. The scheme of the polymerization process is shown in Fig. 1.

Table 1 Operational condition employed for each polymerization reaction

Operational condition
Reaction	Ratio of (S : DVB)	Comonomer feed flow rate (L h^−1)
1	3 : 1	—
2	3 : 1	0.032
3	3 : 1	0.076
4	3 : 1	0.122
5	3 : 1	0.019
6	3 : 1	0.037
7	3 : 1	0.069

---

Fig. 1 Scheme of the production of core–shell particles by simultaneous suspension and emulsion polymerization process.

The morphological characterization of the supports (specific area, average pore diameter and volume of pores) was determined by nitrogen physisorption, using a surface analyzer (ASAP 2020 model, supplied by Micromeritics, Norcross, GA, USA) and the obtained values were adjusted using the BET model. Sample treatment was performed under vacuum at 60 °C. The average particle diameters were evaluated with a particle size distribution analyzer supplied by Malvern Instruments (Master sizer Hydro 2000S model). Measurements were performed in duplicates and the experimental errors were calculated with confidence level of 95%.

Polymer particles were also characterized by optical microscopy. The binocular microscope (Nikon, model SMZ800 with capacity of 50× magnification) was equipped with a digital camera (Nikon Coolpix 995), enabling the amplification and digitization of the images. A Scanning Electron Microscope (Fei Company, Model Quanta 200) was also used to characterize the obtained particles. Photomicrographs were processed in an image analyzer (Fei Company). Typical morphological features of PS/PS particles have been described in previous publications.⁶⁰

2.2.1. Hydrolytic activity by p-NPB method and protein determination. The hydrolytic activities of the free or immobilized enzymes were determined by measuring the increase of absorbance at 348 nm (isobestic point of pNP, ε is 5150 M⁻¹ cm⁻¹)⁶³ produced by the release of pNP during the hydrolysis of 0.4 mM p-NPB in 50 mM sodium phosphate at pH 7.0 and 25 °C. The reaction was initialized by adding 50–100 μL of the lipase solution or suspension to 2.5 mL of the substrate solution. One international unit of activity (U) was defined as the amount of enzyme that hydrolyzes 1 μmol of p-NPB per minute under the previously described conditions. Protein concentration was determined following Bradford's method,⁶⁴ using bovine serum albumin as the reference.

2.2.2. Wetting of the different supports. PS/PS and PS-co-DVB/PS-co-DVB supports were pretreated to facilitate wetting. 1 g of support was suspended in 10 mL of ethanol under slow stirring for 30 min, in order to fill the support of solvent to remove air from the pores of the supports, facilitating the penetration of the enzymatic solution. Afterwards, the support particles were filtrated, and suspended in 10 mL of distilled water. The suspension was maintained under stirring for 30 min. Finally, the particles were filtrated and washed with abundant distilled water.

2.2.3. Immobilization of lipases on core–shell and octyl-agarose beads supports. CALB, TLL, RML and LU were immobilized onto the core–shell supports and octyl agarose via interfacial activation. Standard immobilizations were performed using 1 mg of protein per g of treated support to prevent diffusional constraints. To determine protein loading capacity, the amount of enzyme was increased continuously until reduction of the immobilization yield. In this case, enzyme immobilization was considered to be complete when no significant changes of the supernatant activity could be detected after 4 h. Initially, commercial enzyme solutions were diluted in 10 mM of sodium phosphate at pH 7 and 25 °C to give the desired enzyme concentrations. Then, the supports were added to the diluted enzyme solutions. The activities of both supernatant and suspension were measured using p-NPB. At the end of the immobilization process, suspensions were filtrated and the immobilized enzymes were washed several times with distilled water.

2.2.4. Study of the stability of the different biocatalysts.
2.2.4.1. Thermal inactivation of different immobilized enzymes. 1 g of immobilized enzyme was suspended in 10 mL of 25 mM of sodium phosphate at pH 7 at different temperatures to force enzyme inactivation (reaction conditions were selected in order to obtain inactivation rates based on reliable data in a reasonable time). Then, samples were withdrawn and the enzymatic activity was measured using the p-NPB method described above. The half-lives were calculated from the observed inactivation courses.
2.2.4.2. Inactivation of different immobilized enzymes in the presence of organic co-solvents. Biocatalysts were incubated in mixtures of 90% (v/v) DMF (dimethylformamide) and 10% (v/v) of 100 mM Tris–HCl. The pH of the inactivating solution was adjusted at 7 before adding the immobilized enzymes at 4 °C. Then, the temperature was set to 25 °C. Then, samples were periodically withdrawn and the enzymatic activity was measured with the p-NPB assay. The half-lives were calculated from the observed inactivation courses. The addition of 90 μL of DMF during the activity tests did not affect the observed activity.

2.2.5. Hydrolysis of methyl mandelate. Enzymatic activity was also evaluated using methyl mandelate as substrate and the respective maximum loaded biocatalysts. 1 g of immobilized enzyme was added to 10 mL of 50 mM methyl mandelate dissolved in 50 mM sodium phosphate at pH 7 and 25 °C under continuous stirring. Substrate conversions were determined using a RP-HPLC (Spectra Physic SP 100) coupled with an UV detector (Spectra Physic SP 8450) and a Kromasil C18 (15 cm × 0.46 cm) column (Análisis vinicos, Spain). During analysis, 20 μL of each sample were injected in the column at a flow rate of 1.0 mL min⁻¹ utilizing a solution of acetonitrile: 10 mM ammonium acetate (35 : 65, v/v) (pH = 2.8) as mobile phase and the absorbance at 230 nm was recorded (retention times were 2.4 min for mandelic acid and 4.2 min for methyl mandelate). One unit of enzyme activity was defined as the amount of enzyme necessary to produce 1 μmol of mandelic acid per minute under the conditions described above. Each reaction was executed in triplicate with a maximum conversion of 15–20%. Reported data are based on average values.

2.2.6. Hydrolysis of triacetin. Maximum loaded biocatalysts were also used in the hydrolysis of triacetin. Solutions of 100 mM triacetin in 100 mM sodium phosphate at pH 7 were prepared. 1 g of the biocatalyst was added to 50 mL of the substrate solution. Reactions were performed under stirring at 25 °C. Samples were periodically withdrawn from reaction suspensions. The biocatalyst was discarded by centrifugation and the concentration of products in the supernatant was analyzed by HPLC. A solution of 10% acetonitrile/water (v/v) was used as the mobile phase and a Kromasil C18 column (15 cm × 0.46 cm) was employed. A RP-HPLC (Spectra Physic SP 100) coupled with an UV detector Spectra Physic SP 8450 (detection was performed at 230 nm) were used (retention times of 32.0 min for triacetin, 5.8 min for 1,2-diacetin and 4.8 min for 1,3-diacetin).⁶⁵ Concentrations of triacetin were calculated based on calibration curves using real samples. Reactions were performed in triplicates with maximum conversions of 15–20% and the reported data are based on the mean of the obtained values.

3. Results and discussion

3.1. Influence of the polymerization conditions on the morphology of the supports

Fig. 2 shows that the increase in the co-monomer feed flow rate causes a decrease in the average particle diameter. Probably higher co-monomer feed flow rates destabilize the emulsion media, increasing the agglomeration of the emulsified nanoparticles thus diminishing the average core–shell particle diameter. However, the average particle diameters are still on the micrometric scale even using high flow rates.

Fig. 2 Influence of the co-monomer feed flow rate on average particle diameter.

Fig. 3 illustrates the influence of the co-monomer feed flow rate on the morphology of the polymeric particles. Considering the specific area (Fig. 3A) and the volume of pores (Fig. 3B) of the core–shell particles, there was a particular range on the monomer feed flow rate that resulted in particles with pronounced specific area and porosity. However, apparently the comonomer feed flow rate did not affect the average pore diameter of the particles (Fig. 3C). Probably, a low co-monomer feed flow rate (0.019 L h⁻¹) provided a longer time for the coating of the cores (that could result in higher specific area and more porous particles). However, it may also cause greater stability of the emulsified particles, and that may result in lower core coatings and lower specific area of the core–shell particles. Moreover, a high co-monomer feed flow rate caused a destabilization of the emulsion, resulting in larger agglomeration of particles (both the core and the shell nanoparticles) and in a decrease of the specific area and the porosity. Therefore, regarding the enzyme immobilization procedure, only one support among the ones that were obtained was evaluated: the one having the highest specific area (produced on reaction 7), called PS-co-DVB/PS-co-DVB. The support PS/PS was used for comparison.

Fig. 3 Effect of the comonomer feed flow rate on the morphological properties of the particles: (A) specific area; (B) volume of pores; (C) average pore diameter.

3.2. Morphological aspects of the synthesized supports employed on enzyme immobilization

The morphological aspect of PS-co-DVB/PS-co-DVB and PS/PS core–shell particles and also the PS-co-DVB core particles are illustrated in Fig. 4. The compact PS-co-DVB core particles were much smaller than the core–shell particles and exhibited a well-defined spherical morphology. The PS/PS core–shell particles were larger and presented characteristic irregular surfaces. Comparing both types of core–shell particles, it can be noted that the new PS-co-DVB/PS-co-DVB particles showed much more regular spherical appearance and were smaller than PS/PS particles. Fig. 5 shows that core–shell particles become more irregular and much larger when the emulsified particles agglomerate over the cores to form the shell structure. Formation of the porous shell is clearly visualized in Fig. 5C.

Fig. 4 Optical micrographs of the produced polymeric supports (the length of the ruler is equivalent to 100 μm): (A) core–shell PS-co-DVB/PS-co-DVB particles; (B) core–shell PS/PS particles; (C) compact PS-co-DVB particles.

Fig. 5 Scanning electron micrographs of the polymeric particles: (A) PS-co-DVB core particles; (B) core–shell PS/PS particles; (C) core–shell PS-co-DVB/PS-co-DVB particles.

Considering the wide range of likely applications for the produced supports, the average particles diameter is important since it may condition their handling. Very small particles would require the use of complex methods for separation of biocatalysts from the reaction media at the end. Very large particles can intensify diffusional problems. Therefore, the production of micrometric support particles is desirable and can facilitate the industrial use of the biocatalysts. The average particle diameters of the supports that will be used on the enzyme immobilization process are shown in Table 2. The produced supports presented average particle diameters of approximately 100 μm. PS/PS particles were the largest ones. It can also be noticed that the average diameters of the new PS-co-DVB/PS-co-DVB core–shell particles were smaller than those of the corresponding core particles. This was due to coagulation of nanoparticles during the emulsion step, which shifted the average particle sizes towards smaller values. Nevertheless, the particles were still on the micrometric scale, which is advantageous for separation processes.

Table 2 Average diameters of the produced particles

Supports	Average particle diameter (d50) (μm)
a Core–shell supports.b The compact cores particles alone.
PS/PS^a	114.6 ± 1.3
PS-co-DVB/PS-co-DVB^a	65.8 ± 18.6
PS-co-DVB^b	92.2 ± 1.1

---

Table 3 shows the specific area, the average pore diameter and the volume of pores of the produced core–shell particles. These results clearly indicate the formation of the shell over the particle cores, as the core presents negligible specific area when compared to the core–shell supports.

Table 3 Morphological characteristics of the produced supports used on enzyme immobilization

Supports	Specific area (m^2 g^−1)	Average pore diameter (Å)	Specific volume of pores (m^3 g^−1)
a Pinto et al., 2014.^60b Core–shell support.c The compact core particles alone.
PS/PS^a^,^b	27.3	287.6	0.20
PS-co-DVB/PS-co-DVB^b	33.4	197.2	0.16
PS-co-DVB^c	0.0025	—	—

---

Comparing the core–shell supports, the specific area and porosity were higher for the new PS-co-DVB/PS-co-DVB particles, indicating that the small modifications of the implemented operation procedures utilized in the preparation of these new materials allowed the production of more porous matrices. The presence of DVB in the reaction media leads to an increase in the particle porosity because DVB promotes chain crosslinking, changing the microstructure of polymeric particles, as discussed in previous works.⁴¹ Besides, the increase of the feed flow rate during the emulsion step (0.069 L h⁻¹ instead of 0.04 L h⁻¹, as in the previous study⁴¹) allowed the production of higher amounts of nanoparticles in shorter reaction times, increasing the desired core coverage without increasing the rate of undesired agglomeration of the support particles.

It is possible to observe in Table 3 that the average pore sizes of the supports ranged from 200 Å (PS-co-DVB/PS-co-DVB) to 290 Å (PS/PS), which are wide enough to permit the diffusion of even moderately large protein molecules. As the core compact particles exhibited negligible specific areas, they were not used for enzyme immobilization studies; they were used only to evaluate the shell coverage on the core–shell particles synthesis.

According to Cunha et al. (2014)⁴¹ all of these different supports should exhibit similar hydrophobicities. This indicates that distinct interactions between the polymeric supports and the enzymes should be mainly caused by differences of the surface characteristics of the particles (area, internal morphology). Moreover, as the synthesized supports are hydrophobic, lipases should tend to become interfacially activated versus their surface, and this should be the main cause for the immobilization procedure.²⁹

3.3. Performance of the different supports on the immobilization of the different enzymes

3.3.1. Effects on the activity. Fig. 6 shows the immobilization courses using low enzyme loading (10–20 pNPB U) to prevent diffusional problems that could alter the results. Immobilization using these low loadings is almost total and very rapid for all assayed supports and enzymes. Moreover, all three supports allowed the increase of the enzymes activity upon immobilization, ranging from 109% (octyl-agarose-CALB) to 377% (octyl-agarose-RML). This increase in enzyme activity has been previously reported and derived from the stabilization of the open form of the lipases.³³ In most cases, the new PS-co-DVB/PS-co-DVB was the support that gave the highest activity (except for RML, in this case octyl-agarose gave the highest activity), and PS/PS gave the lowest activities (except for CALB, where octyl-agarose-CALB gave the lowest activity).

Fig. 6 Immobilization courses of different enzymes (10 U) onto different supports: (A) CALB; (B) RML; (C) TLL; (D) LU. PS-co-DVB/PS-co-DVB supports are represented by squares; PS/PS supports are represented by triangles; octyl-agarose supports are represented by circles. Dotted lines represent the suspension activities; continuous lines show the supernatant activities.

3.3.2. Determination of the loading capacity of the different core–shell supports. The immobilization yields at different protein concentrations of CALB, RML, LU and TLL on the different supports are illustrated in Fig. 1S–3S (ESI†). The new PS-co-DVB/PS-co-DVB particles immobilized the activity contained in 65 mg g⁻¹ of CALB, 120 mg g⁻¹ of RML, 55 mg g⁻¹ of LU and 65 mg g⁻¹ of TLL (Fig. 1S(A–D)†). When PS/PS supports were employed, the maximum enzymatic loadings were lower regardless the analyzed enzyme (around 55–70% of the results obtained using PS-co-DVB/PS-co-DVB core–shell particles, depending on the lipase) (Fig. 2S(A–D)†). This result reflected the lower specific areas of PS/PS (around 80% of the specific area of PS-co-DVB/PS-co-DVB particles), and the success in the new design of the core–shell process. The differences between loading capacity and specific area may be due to changes of both supports under wet conditions, as apparently PS/PS supports took 20% more water than PS-co-DVB/PS-co-DVB, which can explain the differences because we used wet supports weights in the immobilization experiments. Moreover, the dynamics of the polymer chains can be very different in both supports, as the copolymer chains crosslinked by DVB became more rigid and this may somehow affect the final support performance. Octyl-agarose was able to immobilize around 20 mg g⁻¹ of CALB, 15 mg g⁻¹ of RML 22 mg g⁻¹ of TLL and 27 mg g⁻¹ using LU (Fig. 3S†). Thus, both core–shell supports showed higher enzyme loading capacities than octyl agarose, which is considered a very good support for lipase immobilization.⁵⁰

3.3.3. Stability of the different biocatalysts under different conditions. Table 4 shows the half-lives of the different immobilized enzymes when subjected to thermal inactivation conditions at pH 7 and in the presence of DMF. It is important to point out that the most stable biocatalysts depended on the enzyme nature and on the inactivation conditions; there was not a “universal” optimal support for the four enzymes. PS-co-DVB/PS-co-DVB-CALB was more stable in thermal inactivation than PS/PS-CALB but it was less stable in presence of DMF. The results using TLL and LU biocatalysts showed the opposite behavior. On the other hand, PS-co-DVB/PS-co-DVB-RML biocatalyst was always more stable than the other core–shell preparation. Octyl-agarose-CALB (a support that has been reported that may stabilize lipases hundreds folds compared to the free lipase)⁶⁶ exhibited a thermal stability that was similar to the thermal stability of PS-co-DVB/PS-co-DVB-CALB, but it was significantly less stable in DMF medium. Octyl-agarose-TLL was the most stable TLL biocatalyst under thermal inactivation conditions, but was less stable than PS-co-DVB/PS-co-DVB-TLL under solvent inactivation conditions. Finally, PS-co-DVB/PS-co-DVB-LU presented the lowest thermal stability but it was the most stable under solvent inactivation conditions. The lipase preparations were also inactivated at pH 5 and 9, the differences in the stabilities between the different biocatalysts were maintained (see Table 1S†).

Table 4 Half-lives (expressed in minutes) of the different biocatalysts under different inactivation conditions. Experiments were performed as described in Section 2

Lipases	Supports	Inactivation conditions^a
		pH 7, 70 °C	90% DMF, pH 7, 25 °C
a The number of replicates of these analyses were 6 (n = 6) and the experimental errors were calculated with confidence level of 95%.
CALB	PS/PS	10.0 ± 0.5	45 ± 1.0
	PS-co-DVB/PS-co-DVB	30 ± 0.5	25 ± 1.0
	Octyl-agarose	28 ± 1.5	8 ± 1.0
RML	PS/PS	5.0 ± 0.2	5.0 ± 1.0
	PS-co-DVB/PS-co-DVB	20 ± 0.2	18 ± 0.8
	Octyl-agarose	10 ± 0.5	13 ± 2.0
TLL	PS/PS	120 ± 1.0	20 ± 1.0
	PS-co-DVB/PS-co-DVB	40 ± 0.1	40 ± 0.5
	Octyl-agarose	165 ± 5.0	22 ± 3.0
LU	PS/PS	40 ± 0.5	15 ± 1.0
	PS-co-DVB/PS-co-DVB	5 ± 0.2	40 ± 1.0
	Octyl-agarose	19 ± 4.0	8 ± 1.5

---

Therefore, the relative stability of the preparations depends on the enzyme and also on the inactivating agent, although, in general, in the presence of organic co-solvents the better performance of the new core–shell biocatalysts seems clear. Perhaps the high hydrophobicity of the produced particles permits a stronger enzyme adsorption on the support, when compared to octyl-agarose biocatalysts. This may reduce the release of the enzyme to the medium in the presence of organic solvents, being the main reason for lipase inactivation in organic solvent during incubation in high organic cosolvent concentration solutions.⁶

3.3.4. Enzyme activity versus different ester substrates. All biocatalysts were evaluated in hydrolysis reactions of two compounds with very different structures, triacetin and methyl mandelate (the R and the S isomers) (Tables 5 and 6, respectively). These assays were performed in order to check if the specificity of the enzyme may be altered after immobilization on supports having different properties even if the mechanism of immobilization is similar, as observed in other works.⁶

Table 5 Activity of different enzyme biocatalysts in the hydrolysis of triacetin. Experimental conditions are described in Section 2

	Hydrolysis of triacetin
	Biocatalyst	Activity (μmol (min g)^−1)
a Analyses were conducted in triplicate (n = 3) and the experimental errors were calculated with confidence level of 95%.
CALB	Novozym 435	207 ± 1.2
	PS/PS	214 ± 1.0
	PS-co-DVB/PS-co-DVB^a	105 ± 0.8
	Octyl-agarose	89 ± 0.5
RML	RM-IM	28.1 ± 0.5
	PS/PS	47.0 ± 0.3
	PS-co-DVB/PS-co-DVB	50.0 ± 0.6
	Octyl-agarose	42.8 ± 0.5
TLL	TL-IM	15.1 ± 0.2
	PS/PS	32.1 ± 0.6
	PS-co-DVB/PS-co-DVB	32.2 ± 0.5
	Octyl-agarose	27.3 ± 1.1
LU	PS/PS	20.8 ± 1.1
	PS-co-DVB/PS-co-DVB	31.3 ± 0.8
	Octyl-agarose	28.3 ± 0.5

---

Table 6 Activity of different enzyme biocatalysts in the hydrolysis of R methyl mandelate and VR/VS ratio. Experimental conditions are described in Section 2^a

	Hydrolysis of R-methyl mandelate
	Biocatalyst	Activity (μmol (min g)^−1)	VR/VS
a Analyses were conducted in triplicate (n = 3) and the experimental errors were calculated with confidence level of 95%.
CALB	Novozym 435	140 ± 1.2	1.35
	PS/PS	210 ± 2.2	1.65
	PS-co-DVB/PS-co-DVB	260 ± 1.6	2.65
	Octyl-agarose	210 ± 2.8	2.30
RML	RM-IM	0.51 ± 0.3	0.32
	PS/PS	0.57 ± 0.6	0.31
	PS-co-DVB/PS-co-DVB	0.82 ± 0.1	0.28
	Octyl-agarose	0.74 ± 0.1	0.24
TLL	TL-IM	1.0 ± 0.2	2.3
	PS/PS	1.1 ± 0.1	1.64
	PS-co-DVB/PS-co-DVB	2.3 ± 0.1	1.67
	Octyl-agarose	1.1 ± 0.2	1.57
LU	PS/PS	0.9 ± 0.2	0.81
	PS-co-DVB/PS-co-DVB	2.6 ± 0.2	1.45
	Octyl-agarose	0.95 ± 0.1	0.95

---

Using CALB biocatalysts in the hydrolysis of triacetin, the least active biocatalyst was octyl-agarose-CALB followed by PS-co-DVB/PS-co-DVB-CALB. PS/PS-CALB was slightly more active than the commercial and widely used Novozym 435 with this substrate (Table 5). However, in the hydrolysis of R methyl mandelate, Novozym 435 had the lowest activity while the new PS-co-DVB/PS-co-DVB-CALB was the most active, with octyl-agarose-CALB and PS/PS-CALB exhibiting similar activities (Table 6). All preparations preferred the R isomer with a moderate enantiopreference, although PS-co-DVB/PS-co-DVB-CALB doubled the ratio VR/VS compared to Novozym 435.

Analyzing RML biocatalysts, the most active one using both substrates was the new PS-co-DVB/PS-co-DVB-RML (Tables 5 and 6). The second most active biocatalyst depended on the employed substrate: it was PS/PS-RML using triacetin, while using methyl mandelate as substrate, the second most active was octyl-agarose-RML. The commercial preparation was the catalyst with the lowest activity in both cases. In this case, the preferred isomer was the S, with activity ratios between both isomers ranging from 3 to 4 (Table 6).

Considering TLL immobilized preparations, the most active biocatalysts in triacetin hydrolysis were both core–shell supports, shortly followed by octyl-agarose-TLL and both of them doubling the activity of the commercial preparation (Table 5). Using R methyl mandelate, PS-co-DVB/PS-co-DVB-TLL was the most active; octyl-agarose-TLL and PS/PS-TLL showed around half of that activity, shortly followed by the commercial preparation (Table 6). The preference for the R isomer was scarce (ranging from 1.57 using octyl-agarose-TLL to 2.3 employing the commercial preparation).

LU biocatalysts exhibited short differences using triacetin as substrate, being PS-co-DVB/PS-co-DVB-LU the most and PS/PS the least active one (Table 5). However, PS-co-DVB/PS-co-DVB-LU was 2.5 fold more active than the other two biocatalysts using R methyl mandate as substrate, (Table 6). The enantiopreference was very low, but while PS-co-DVB/PS-co-DVB-LU preferred the R isomer, the other two preparations preferred the S isomer (Table 6).

It was possible to observe generally better hydrolytic activities of the new biocatalysts, mainly the enzymes immobilized in the new PS-co-DVB/PS-co-DVB core–shell support, compared to the commercial ones. More interestingly, it was noticed that even though the immobilization of all home-made biocatalysts involved interfacial activation as immobilization mechanism, the final properties of each one were strongly modulated by the exact nature of the support. Thus, the most active biocatalyst produced using a specific support may exhibit a low hydrolytic activity when other substrate were investigated, as reported in many other cases.⁶

Another important feature of an immobilized enzyme is its operational stability. To this goal, each enzyme biocatalyst was employed on 5 consecutive cycles of hydrolysis of triacetin. After each cycle, the biocatalysts were washed 3 times with 3 volumes of 20 mM sodium phosphate. All the biocatalysts, including the commercial biocatalysts, those prepared using octyl-agarose and core–shell biocatalysts, showed a decrease of activity under 20% along the 5 reaction cycles, as shown in Fig. 7.

Fig. 7 Re-use of different enzyme biocatalysts on hydrolysis of triacetin: (A) CALB; (B) RML; (C) TLL; (D) lecitase. Rhombi: commercial preparations; PS-co-DVB/PS-co-DVB: triangles; PS/PS: squares.

4. Conclusion

The new hydrophobic core–shell support (PS-co-DVB/PS-co-DVB) developed in the present work showed high protein loading capacity, exceeding the capacity of commercial octyl-agarose supports, known for its very good performance³³ and that of the previously described PS/PS core shell.⁴¹ This high enzyme loading capacity is obtained by the porous shell of the produced polymeric core/shell particles, as we have shown that the core loading capacity is negligible. However, the core particles are critical to produce stable particles.⁶¹

The new biocatalysts were much more stable than the commercial biocatalysts or the ones obtained using octyl-agarose in many instances but not always. This higher stability was mainly observed in organic solvents inactivations, where enzyme desorption play an important role in the biocatalyst stability and the more hydrophobic nature of the new PS-co-DVB/PS-co-DVB should reduce the enzyme release.⁶⁷ However, considering each enzyme and each inactivation condition, it seems that there is not an absolute “optimal” immobilization support from those here studied regarding enzyme stability.

Finally, the activities of the new biocatalysts employing distinct substrates were much higher than the ones obtained with commercial products, except when triacetin was hydrolyzed by CALB. Moreover, in many cases, the enzymatic activities were also much higher than the ones observed when octyl-agarose was used as support. As it has been previously reported,³⁵ the chemical and textural properties of the support surfaces alter the final lipase performance even if the immobilization mechanism is in all cases interfacial activation. Thus, the use of differently prepared hydrophobic core–shell supports may be a way to enrich a library of lipase biocatalysts.⁸

Based on the obtained results, it can be concluded that changing some operational conditions during the polymerization reaction, such as the co-monomer feed flow rate, it is possible to synthesize core–shell particles with distinct morphological aspects. Moreover, among the different core–shell supports that were produced, the new PS-co-DVB/PS-co-DVB constitutes a very promising support for lipase immobilization.

Acknowledgements

The authors thank CNPq, CAPES and FAPERJ for the scholarships, for financial support and for funding the project PVE 301139/2014-8. The authors gratefully recognize the support from the MINECO of Spanish Government by the grants CTQ2013-41507-R and CTQ2016-78587-R. The authors also thank Mr Ramiro Martínez (Novozymes, Spain) for the kind supply of the enzymes used in the present research. The help and comments from Dr Ángel Berenguer (Instituto de Materiales, Universidad de Alicante) are kindly acknowledged.

References

1. I. Chibata, T. Tosa and T. Sato, J. Mol. Catal., 1986, 37, 1–24.
2. W. Hartmeier, Trends Biotechnol., 1985, 3, 149–153.
3. E. Katchalski-Katzir, Trends Biotechnol., 1993, 11, 471–478.
4. J. F. Kennedy, E. H. M. Melo and K. Jumel, Chem. Eng. Prog., 1990, 86, 81–89.
5. R. Fernandez-Lafuente, Enzyme Microb. Technol., 2009, 45, 405–418.
6. C. Mateo, J. M. Palomo, G. Fernandez-Lorente, J. M. Guisan and R. Fernandez-Lafuente, Enzyme Microb. Technol., 2007, 40, 1451–1463.
7. M. Petkar, A. Lali, P. Caimi and M. Daminati, J. Mol. Catal. B: Enzym., 2006, 39, 83–90.
8. R. C. Rodrigues, C. Ortiz, Á. Berenguer-Murcia, R. Torres and R. Fernández-Lafuente, Chem. Soc. Rev., 2013, 42, 6290–6307.
9. K. Hernandez and R. Fernandez-Lafuente, Enzyme Microb. Technol., 2011, 48, 107–122.
10. U. Guzik, K. Hupert-Kocurek and D. Wojcieszyńska, Molecules, 2014, 19, 8995–9018.
11. C. Garcia-Galan, Á. Berenguer-Murcia, R. Fernandez-Lafuente and R. C. Rodrigues, Adv. Synth. Catal., 2011, 353, 2885–2904.
12. O. Barbosa, C. Ortiz, A. Berenguer-Murcia, R. Torres, R. C. Rodrigues and R. Fernandez-Lafuente, Biotechnol. Adv., 2015, 33, 435–456.
13. C. Carboni-Oerlemans, P. Domínguez de María, B. Tuin, G. Bargeman, A. Van der Meer and R. J. Van Gemert, Biotechnology, 2006, 126, 140–151.
14. M. S. Humble and P. Berglund, Eur. J. Org. Chem., 2011, 3391–3401.
15. M. Kapoor and M. N. Gupta, Process Biochem., 2012, 47, 555–569.
16. K. E. Jaeger and T. Eggert, Curr. Opin. Biotechnol., 2002, 13, 390–397.
17. P. Adlercreutz, Chem. Soc. Rev., 2013, 42, 6406–6436.
18. R. Sharma, Y. Chisti and U. C. Banerjee, Biotechnol. Adv., 2001, 19, 627–662.
19. K. E. Jaeger and M. T. Reetz, Trends Biotechnol., 1998, 16, 396–403.
20. P. Lozano, Green Chem., 2010, 12, 555–569.
21. P. Lozano, E. Garcia-Verdugo, S. Luis, M. Pucheault and M. Vaultier, Curr. Org. Synth., 2011, 8, 810–823.
22. R. Verger, Trends Biotechnol., 1997, 15, 32–38.
23. C. Cambillau, S. Longhi, A. Nicolas and C. Martinez, Curr. Opin. Struct. Biol., 1996, 6, 449–455.
24. A. M. Brzozowski, U. Derewenda, Z. S. Derewenda, G. G. Dodson, D. M. Lawson and J. P. Turkenburg, et al., Nature, 1991, 351, 491–494.
25. K. K. Kim, H. K. Song, D. H. Shin, K. Y. Hwang and S. W. Suh, Structure, 1997, 5, 173–185.
26. K. E. Jaeger, S. Ransac, H. B. Koch, F. Ferrato and B. W. Dijkstra, FEBS Lett., 1993, 332, 143–149.
27. M. Cygler and J. D. Schrag, Biochim. Biophys. Acta, 1999, 1441, 205–214.
28. P. Reis, K. Holmberg, H. Watzke, M. E. Leser and R. Miller, Adv. Colloid Interface Sci., 2009, 147–148, 237–250.
29. R. Fernandez-Lafuente, P. Armisén, P. Sabuquillo, G. Fernández-Lorente and J. M. Guisán, Chem. Phys. Lipids, 1998, 93, 185–197.
30. J. M. Palomo, C. Ortiz, G. Fernández-Lorente, M. Fuentes, J. M. Guisán and R. Fernández-Lafuente, Enzyme Microb. Technol., 2005, 36, 447–454.
31. G. Fernández-Lorente, J. M. Palomo, M. Fuentes, C. Mateo, J. M. Guisán and R. Fernández-Lafuente, Biotechnol. Bioeng., 2003, 82, 232–237.
32. J. M. Palomo, M. M. Peñas, G. Fernández-Lorente, C. Mateo, A. G. Pisabarro and R. Fernández-Lafuente, et al., Biomacromolecules, 2003, 4, 204–210.
33. E. A. Manoel, C. S. dos Santos, D. M. G. Freire and R. Fernandez-Lafuente, Enzyme Microb. Technol., 2015, 71, 53–57.
34. Z. Cabrera, G. Fernandez-Lorente, R. Fernandez-Lafuente, J. M. Palomo and J. M. Guisan, J. Mol. Catal. B: Enzym., 2009, 57, 171–176.
35. G. Fernandez-Lorente, Z. Cabrera, C. Godoy, R. Fernandez-Lafuente, J. M. Palomo and J. M. Guisan, Process Biochem., 2008, 43, 1061–1067.
36. R. C. Rodrigues, C. Garcia-Galan, O. Barbosa, K. Hernandez, J. C. S. dos Santos, A. Berenguer-Murcia and R. Fernandez-Lafuente, Curr. Org. Chem., 2015, 19, 1707–1718.
37. E. Taqieddin and M. Amiji, Biomaterials, 2004, 25, 1937–1945.
38. K. M. Ho, X. Mao, L. Gu and P. Li, Langmuir, 2008, 24, 11036–11042.
39. Q. Gu, Q. Lin, C. Hu and B. Yang, J. Appl. Polym. Sci., 2005, 95, 404–412.
40. M. D. Besteti, D. M. G. Freire and J. C. Pinto, Macromol. React. Eng., 2011, 5, 518–532.
41. A. G. Cunha, M. D. Besteti, E. A. Manoel, A. A. T. Da Silva, R. V. Almeida and A. B. C. Simas, et al., J. Mol. Catal. B: Enzym., 2014, 100, 59–67.
42. R. C. Rodrigues and R. Fernandez-Lafuente, J. Mol. Catal. B: Enzym., 2010, 64, 1–22.
43. R. Fernandez-Lafuente, J. Mol. Catal. B: Enzym., 2010, 62, 197–212.
44. E. M. Anderson, K. M. Larsson and O. Kirk, Biocatal. Biotransform., 1998, 16, 181–204.
45. J. Uppenberg, M. T. Hansen, S. Patkar and T. A. Jones, Structure, 1994, 2, 293–308.
46. Z. S. Derewenda, U. Derewenda and G. G. Dodson, J. Mol. Biol., 1992, 227, 818–839.
47. U. Derewenda, L. Swenson, R. Green, Y. Wei, S. Yamaguchi and R. Joerger, et al., Protein Eng., 1994, 7, 551–557.
48. K. Clausen, Eur. J. Lipid Sci. Technol., 2001, 103, 333–340.
49. L. De Maria, J. Vind, K. M. Oxenbøll, A. Svendsen and S. Patkar, Appl. Microbiol. Biotechnol., 2007, 74, 290–300.
50. A. Bastida, P. Sabuquillo, P. Armisen, R. Fernández-Lafuente, J. Huguet and J. M. Guisán, Biotechnol. Bioeng., 1998, 58, 486–493.
51. J. C. S. Dos Santos, N. Rueda, R. Torres, O. Barbosa, L. R. B. Gonçalves and R. Fernandez-Lafuente, Process Biochem., 2015, 50, 918–927.
52. J. C. S. Dos Santos, N. Rueda, A. Sanchez, R. Villalonga, L. R. B. Gonçalves and R. Fernandez-Lafuente, RSC Adv., 2015, 5, 35801–35810.
53. K.-E. Jaeger, S. Ransac, H. B. Koch, F. Ferrato and B. W. Dijkstra, FEBS Lett., 1993, 332, 143–149.
54. M. Cygler and J. D. Schrag, BBA, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids, 1999, 1441, 205–214.
55. J. M. Palomo, M. Fuentes, G. Fernández-Lorente, C. Mateo, J. M. Guisan and R. Fernández-Lafuente, Biomacromolecules, 2003, 4, 1–6.
56. G. Fernández-Lorente, J. M. Palomo, M. Fuentes, C. Mateo, J. M. Guisán and R. Fernández-Lafuente, Biotechnol. Bioeng., 2003, 82, 232–237.
57. L. Wilson, J. M. Palomo, G. Fernández-Lorente, A. Illanes, J. M. Guisán and R. Fernández-Lafuente, Enzyme Microb. Technol., 2006, 39, 259–264.
58. J. M. Palomo, C. Ortiz, G. Fernández-Lorente, M. Fuentes, J. M. Guisán and R. Fernández-Lafuente, Enzyme Microb. Technol., 2005, 36, 447–454.
59. J. M. Palomo, C. Ortiz, M. Fuentes, G. Fernandez-Lorente, J. M. Guisan and R. Fernandez-Lafuente, J. Chromatogr. A, 2004, 1038, 267–273.
60. M. C. C. Pinto, D. M. G. Freire and J. C. Pinto, Molecules, 2014, 19, 12509–12530.
61. M. D. Besteti, A. G. Cunha, D. M. G. Freire and J. C. Pinto, Macromol. Mater. Eng., 2014, 299, 135–143.
62. M. K. Lenzi, F. M. Silva, E. L. Lima and J. C. Pinto, J. Appl. Polym. Sci., 2003, 89, 3021–3038.
63. D. Lombardo and O. Guy, Biochim. Biophys. Acta, 1981, 657, 425–437.
64. M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
65. K. Hernandez, E. Garcia-Verdugo, R. Porcar and R. Fernandez-Lafuente, Enzyme Microb. Technol., 2011, 48, 510–517.
66. J. C. S. Dos Santos, N. Rueda, A. Sanchez, R. Villalonga, L. R. B. Gonçalves and R. Fernandez-Lafuente, RSC Adv., 2015, 5, 35801–35810.
67. N. Rueda, J. C. S. Dos Santos, R. Torres, C. Ortiz, O. Barbosa and R. Fernandez-Lafuente, RSC Adv., 2015, 5, 11212–11222.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra13350a

‡ These coauthors have contributed equally to this paper.

---