Kinetics-bolstered catalytic study of a high performance lipase-immobilized nanofiber membrane bioreactor

Abstract

Enzyme-immobilized membrane bioreactors combine biocatalysis with separation, and thus have generated interest among applied researchers. Despite the advantages of enzyme-immobilized membrane bioreactors, some issues associated with the membrane and operation parameters still remain to be addressed in order to achieve scale-up of such systems. We report the fabrication of a biphasic lipase-immobilized nanofiber membrane bioreactor with high catalytic performance. A poly(acrylonitrile-co-acrylic acid) non-woven nanofiber membrane was prepared by electrospinning and lipase from Candida rugosa was covalently coupled to it via an activation process by N-(3-dimethylaminopropyl)-N′-ethyl-carbodiimide hydrochloride/N-hydroxy-succinimide. The influences of membrane diameters and operation variables on bioreactor efficiency were studied using the catalytic hydrolysis of olive oil as the model reaction. Furthermore, a pseudo first order model contributed to the evaluation of lipase performance in a more standardized way. A bioreactor activity of 1.85 × 10⁴ U g⁻¹ was obtained under optimum operation conditions; also, the catalytic system exhibited good operational and recycling stability.

---

1. Introduction

Enzymes are becoming increasingly important for industrial applications due to the fact that they operate under mild working conditions, and they catalyze very specific reactions in a highly efficient manner with limited by-product formation.^1,2 Lipases, a group of enzymes specialized in catalyzing the cleavage of carboxyl ester bonds in tri-, di-, and mono-acylglycerols, have a wide range of applications in food, pharmaceutical, and detergent industries.^3,4 Although most lipases are still used in their soluble forms, increasing attempts aimed at improving the economics of lipase-based technology have been implemented through immobilization. In this method, lipases are often attached or incorporated onto or into inert, insoluble supports to enable enzyme re-use as well as offer better catalytic stability, thus improving both the economics and the viability of the process.^5,6

The two fundamental elements of an enzyme immobilization technology are the immobilization technique and the immobilization support. Lipase has been reported to be immobilized via covalent binding, adsorption, cross-linking and encapsulation. Compared with other immobilization techniques, covalent binding is believed to form the most stable lipase-support attachment.⁷ Among the various immobilization supports such as synthetic resins, biopolymers, inorganic particles and liquid membranes, polymeric membranes have always been of great importance.^8,9 On the one hand, polymer support allows for easy control over material structure and composition to meet the specific requirements of the application. On the other hand, a membrane can be advantageous as an immobilization support for its product separation capabilities along with the potential for biocatalytic conversion. Moreover, the continuous removal of the product can shift the reaction equilibrium towards the product side and thereby increasing the reaction productivity, which is a notable advantage of lipase-immobilized membrane bioreactors.^10–12

The practicalities of using a lipase-immobilized membrane reactor are governed by its productivity, lifetime and stability.¹³ A high standard of membrane reactor quality can only be achieved if both membrane and operation parameters are optimized. In order to be productive, the membrane for enzyme immobilization should have high lipase loading capacity and limited mass transfer resistance. Moreover, it is expected to be prepared easily and at a low price. Nanofiber membranes, which can be simply prepared by an electrospinning process, present a promising choice with a large specific area for high enzyme loading and fine porous structure allowing ready mass accessibility.^14–16 Up to now, although a great variety of literature is available on the diverse nature of lipase-immobilized membrane bioreactors, typically the productivity and stability of the membrane are emphasized while the characterization of enzyme related parameters is omitted, which makes a precise comparison between different techniques difficult.¹⁷ Thus, standardization using a kinetic or thermodynamic description of the immobilized lipase performance in a bioreactor system should prove useful for evaluating the bioreactor and its potential industrial applications.

A biphasic lipase-immobilized nanofiber membrane bioreactor was assembled and characterized in order to better understand the performance of the immobilized enzymes. Poly(acrylonitrile-co-acrylic acid) (PANCAA), a polymer that is easily electrospun and contains acrylic acid groups as the covalent binding sites for lipases, was synthesized and electrospun into nanofiber membranes with well-controlled morphology and fiber diameters. The excellent characteristics which belong to theses membranes include good thermal stability, good mechanical stability and tolerance to most solvents, atmosphere, bacteria and photo irradiation. Lipase from Candida rugosa was covalently coupled onto the nanofiber membrane. The influence of fiber diameters and operating variables (namely the aqueous-phase system, pH, temperature, and inter-fiber transfer limitation) on the performance of this biphasic bioreactor was investigated using olive oil hydrolysis as a model reaction and the corresponding fatty acid production was fitted according to a kinetic model. The characterization of the lipase in the bioreactor is expected to provide a foundation for improving bioreactor systems and inspire a fundamental research of such bioreactors to steer their practical usage.

2. Experimental

2.1 Materials

Poly(acrylonitrile-co-acrylic acid) (PANCAA) with a viscosity-averaged molecular weight of 18.1 × 10⁴ g mol⁻¹ was synthesized by a water-phase precipitation copolymerization process. The molar content of acrylic acid in this copolymer was about 10% by elemental analysis. N-(3-Dimethylaminopropyl)-N′-ethyl-carbodiimide hydrochloride (EDC, analytical grade) and N-hydroxysuccinimide (NHS, biological grade) was purchased from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China) and used without further purification. Lipase (from Candida rugosa) powder (1150 units per mg solid), Bradford reagent and bovine serum albumin (BSA, molecular mass: 67 000 Da) were obtained from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA) and used as received. All other chemicals were of analytical grade and used without further purification.

2.2 Preparation of PANCAA nanofiber membrane

Electrospinning was adopted to prepare PANCAA nanofiber membranes. PANCAA was dissolved using N,N′-dimethylformamide (DMF) at 40 °C into a homogeneous solution. After the air bubbles were removed completely, the solution was injected into a 20 mL plastic syringe with a 1.2 mm inner diameter metal needle connected to a high voltage power supply (GDW-a, Tianjin Dongwen High-voltage Power Supply Plant, China). The solution feed rate was controlled by a micro-infusion pump (WZ-50C2, Zhejiang University Medical Instrument Co. Ltd., China).

Electrospinning conditions affect the nanofiber morphology and diameter. In this work, PANCAA nanofiber membranes with different fiber diameters were obtained by controlling the concentration of polymer solution from 3.0 wt% to 6.0 wt%, and at the same time slightly modulating the potential (8–12 kV) and humidity (40–70%) to obtain uniform and continuous nanofibers. Typically, when the solution concentration was 4.0 wt%, electrospinning was performed at a solution feed rate of 1.0 mL h⁻¹ and a potential of 10 kV, with 15 cm between the needle tip and the earthed aluminum collector and an environment humidity of 60%. All the electrospun nanofiber membranes were dried under vacuum at 60 °C for 24 h to remove residual solvent before use.

2.3 Biphasic lipase-immobilized nanofiber membrane bioreactor

Fig. 1 shows the experimental setup of the biphasic bioreactor. It consisted of two cylindrical compartments with an arrangement for holding the PANCAA nanofiber membrane (membrane diameter: 5.0 cm) between them. A peristaltic pump was used between the two compartments for better control of the flow rate. Lipase was immobilized by an EDC–NHS activation procedure. EDC–NHS solution, washing solutions and lipase solution were successively circulated in the reactor cell by the peristaltic pump. To activate the carboxyl groups on the nanofiber surface, the PANCAA nanofiber membrane was thoroughly washed with de-ionized water and immersed in phosphate buffer solution (PBS, 0.05 M, pH 6.0) for 2 h. Then a volume of 200 mL EDC–NHS solution (the concentration of EDC in PBS 6.0 was 24 mg mL⁻¹ and the molar ratio of EDC to NHS was 1 : 1) was added into the reactor to filter through the membrane at 25 °C for 10 h. For the lipase immobilization process, the activated support was washed several times with PBS 6.0 and then another circulation of 200 mL lipase solution (8 mg mL⁻¹ in PBS, 0.05 M, pH 7.0) was employed for 3 h at 25 °C before washing three times with 100 mL PBS pH 7.0. The washings together with the reaction solution were collected for determination of protein concentration. The amount of enzyme immobilized on the membrane was measured by Bradford assay using Coomassie brilliant blue reagent.¹⁸ BSA was used as the standard to construct the calibration curve. The amount of immobilized protein on the membranes was calculated from the protein mass balance among the initial and final lipase solutions, and the washings. The enzyme loading was defined as the amount of enzyme (mg) per gram of the membrane. Each value was the mean of at least three parallel experiments and the standard deviation was within approximately ±5%.

Fig. 1 Schematic representation of the biphasic lipase-immobilized PANCAA nanofiber membrane bioreactor.

For the continuous hydrolysis process in this bioreactor, a sample of 120 mL substrate emulsion was circulated using a peristaltic pump in the compartment facing the immobilized membrane side. The substrate emulsion was prepared by thoroughly mixing 130 mL olive oil with 400 mL gum arabic solution (11% (w/v) gum arabic powder and 1.25% (w/v) CaCl₂·2H₂O) and stored at 4 °C. In the other compartment, 120 mL aqueous solution at a basic pH was circulated simultaneously. The hydrolysis reaction occurred on the surface of the lipase-immobilized membrane and the temperature of the substrate reservoir or reactor cell was controlled to ±0.5 °C by circulating water through the reservoir or reactor jacket. The hydrolytic products, fatty acid and glycerol, extracted in the aqueous phase were neutralized by continuously adding a 0.05 M NaOH standard solution through an automatic titrator, holding the pH of the aqueous solution constant during the whole process. The consumed volume of the NaOH standard solution was recorded periodically and used for evaluating the reaction rate.

One lipase unit corresponded to the release of 1 μmol fatty acid per hour under the assay conditions. The lipase-immobilized membrane activity was the number of lipase units per gram of membrane, whereas lipase activity was defined as the number of lipase units per milligram of protein. Each data point was the average of at least three parallel experiments, and the standard deviation was within approximately ±5%.

2.4 Non-linear kinetic model for the hydrolysis of olive oil

In order to investigate the kinetic rate constants for the hydrolysis of olive oil by immobilized lipases, a non-linear pseudo first order kinetic model was employed.^19,20

According to the differential equation:

Integrating eqn (1) from the boundary conditions t = 0 to t = t and q_t = 0 to q_t = q_t gives:

or:

q_t = q₁(1 − exp^−k₁t) (3)

which is the integrated rate law for a pseudo first order reaction, where: q₁ is the amount of fatty acid produced at equilibrium; k₁ is the equilibrium rate constant of pseudo first order reaction, min⁻¹.

In this work, PANCAA nanofiber membranes with varied fiber diameters and thickness were installed in the bioreactor to measure the bioreactor efficiency. Also, different operation parameters (aqueous-phase system, pH, temperature) were tested with the resulting hydrolysis conversion measured. The results were fitted according to the pseudo first order model.

2.5 Reusability test of the biphasic lipase-immobilized membrane bioreactor

To evaluate their reusability, the tested lipase-immobilized membranes were washed with PBS (0.05 M, pH 7.0) to remove any residual substrate after the hydrolysis reaction, followed by immersion into fresh catalytic reaction media under the same experimental condition. The hydrolysis reactions were conducted at 25 °C using PBS (0.05 M, pH 7.0) as the aqueous-phase system. The same procedure was repeated up to 6 times. The relative catalytic activities of the immobilized lipases were normalized to the highest activity.

3. Results and discussion

3.1 Effect of fiber diameter on enzyme loading

Traditional electrospinning is a mature technique and has been widely used to fabricate polymer fibers with diameters ranging from several hundreds of micrometers down to tens of nanometers. In this work, PANCAA nanofibers with different diameters were successfully fabricated as shown in Fig. S1.† The electrospun nanofibers were uniform without beads and provided good porosity, ensuring the necessary mechanical strength as well as low diffusion resistance for further processing in the bioreactor. Moreover, the fiber diameters could be manipulated conveniently by adjusting the concentration of polymer solution without sacrificing membrane porosity and fiber uniformity. By measuring the nanofiber diameter, from Fig. S1(a) to S1(d),† the diameters were 156.1 ± 23.3, 200.6 ± 35.0, 271.2 ± 30.3 and 387.1 ± 60.7 nm, respectively. Fig. 2 reveals how diameters influenced the amount of lipase immobilized on the membrane. There was a slight increase of enzyme loading from 26.5 ± 0.87 mg g⁻¹ to 29.2 ± 0.79 mg g⁻¹ when the fiber diameter increased from 156.1 ± 23.3 nm to 200.6 ± 35.0 nm, followed by a sharp decrease when the diameter increased further. Only 14.9 ± 0.71 mg lipase was immobilized per gram of PANCAA nanofiber membrane when the fiber diameter was 387.1 ± 60.7 nm. This decrease can be attributed to the fact that when the nanofiber diameter increased, the specific surface area of the nanofiber membrane reduced, leading to a lower amount of carboxyl groups on the fiber surface. As a result, there were less potential reaction sites for covalent coupling with lipase and thus the enzyme loading decreased.

Fig. 2 Effect of fiber diameter on enzyme loading.

3.2 Effect of fiber diameter on bioreactor efficiency

The performance of the biphasic lipase-immobilized nanofiber membrane bioreactor was evaluated in terms of the fatty acid amount extracted in the aqueous phase. For these experiments, all nanofiber membranes tested had a weight of 4 mg, and the hydrolysis reactions were conducted at 25 °C with PBS (0.05 M, pH 7.0) as the aqueous-phase system. Moreover, to check for errors in the determination of the reaction rate, a blank run was performed in which a membrane without enzyme was used and it was found that the pH of the aqueous phase did not change.

To study how fiber diameters affected the bioreactor efficiency, the endurance of the membranes for the series of reactions in the bioreactor needed to be evaluated. Using the membranes with fiber diameters of 156.1 ± 23.3 and 271.2 ± 30.3 nm as examples, their morphologies after each treatment are displayed in Fig. S2.† The fiber morphology was maintained with clear pores and a smooth fiber surface after the membranes went through activation and lipase immobilization processes. Moreover, after the hydrolysis reaction, although a little fouling occurs, which is virtually inevitable in most membrane processes, the membrane structures remained undamaged. All these phenomena ensured the feasibility of using PANCAA nanofiber membranes with varied fiber diameters for the reactions in this bioreactor.

Fig. 3 shows the production of fatty acid catalyzed by the lipase-immobilized PANCAA membranes with various fiber diameters. The increase of fatty acid quantity with time revealed that the hydrolysis reactions were occurring during the test time. When the fiber diameter was 156.1 ± 23.3, 200.6 ± 35.0, 271.2 ± 30.3 and 387.1 ± 60.7 nm, respectively, the membrane activity corresponding to each diameter was 3.38 × 10⁴, 4.96 × 10⁴, 4.52 × 10⁴ and 4.09 × 10⁴ U g⁻¹, indicating the membrane with a fiber diameter of 200.6 ± 35.0 nm produced the most fatty acid. This result is not surprising because the highest lipase loading occurred when using a membrane with a 200.6 ± 35.0 nm diameter. However, upon further investigation, it was found that the highest lipase activity did not appear on the membrane with a 200.6 ± 35.0 nm fiber diameter: the lipase activity was 1.27 × 10³, 1.70 × 10³, 2.12 × 10³ and 2.74 × 10³ U mg⁻¹ corresponded to the increasing fiber diameter. Also, V_max and K_m were respectively assayed for the free and immobilized lipases. V_max, which defines the highest possible velocity when all the enzyme is saturated with substrate, reflects the intrinsic characteristics of the immobilized enzyme, but may be affected by diffusion constrains. K_m, or apparent K_m, which is defined as the substrate concentration that gives a reaction velocity of 1/2 V_max, reflects the effective characteristics of the enzyme and depends upon both partition and diffusion effects. When the fiber diameter was 200.6 ± 35.0 nm, K_m value was 0.45 mM for the free lipase, while the apparent value was 0.72 mM for the immobilized lipase; the V_max value of 46.4 U mg⁻¹ for the free lipase was found to be higher than that for the immobilized one (37.6 U mg⁻¹). This was either due to the conformational changes of the enzyme resulting in a lower possibility of forming a substrate–enzyme complex, or a less accessibility of the substrate to the active sites of the immobilized enzyme caused by the increased diffusion limitation.

Fig. 3 Effect of fiber diameter on bioreactor efficiency.

From the study above, although the high enzyme loading may help improve the overall membrane activity, it does not always translate to higher enzyme activity and thus may lead to a costly waste of enzymes. This phenomenon can be explained by molecular crowding.²¹ In this case, crowding refers to high concentrations of lipases on the membrane surface. High lipase concentrations offer more chances for protein–protein interactions, which may alter the conformations of lipases and in the worst case, can result in lipase denaturation or active site blocking.²² It is obvious from the discussion above that there is a certain tradeoff between lipase activity and lipase loading, and the membrane activity is a compromise between these two parameters. Thus, in order to maximally utilize the enzyme-immobilized membrane without increases in cost, the production of the nanofiber membrane support should try to provide an optimal balance between enzyme cost-effectiveness and membrane activity.

The pseudo first order model was adopted to investigate the hydrolysis reaction in the bioreactor and the fitting curves are shown in Fig. 3. The validity of this model was assessed based on regression coefficients (R). When the fiber diameter was 156.1 ± 23.3, 200.6 ± 35.0, 271.2 ± 30.3 and 387.1 ± 60.7 nm, respectively, the resulting rate constants k₁ were 1.62 × 10⁻², 9.1 × 10⁻³, 1.24 × 10⁻² and 8.7 × 10⁻³ min⁻¹. The corresponding R values for the pseudo first order model of hydrolysis reactions were 0.994, 0.990, 0.985 and 0.987, all >0.98, indicating the hydrolysis reaction on different fiber diameters followed the pseudo first order model.

3.3 Effect of aqueous-phase system on bioreactor efficiency

Similar membrane conditions guaranteed in this work is a prerequisite when studying the influence of different operation parameters on bioreactor efficiency. Chemical changes of the membrane surfaces were analyzed by X-ray photoelectron spectrum (XPS) to follow the whole lipase immobilization processes, and the chemical composition of the nanofiber membrane surface was shown in Table S1. Fig. 4(a) shows the quantities of fatty acid produced by lipase-immobilized nanofiber membranes in various aqueous-phase systems. During the same time period, the lowest amount of fatty acid was produced when using pure water as the aqueous phase, leading to a membrane activity of only 1.56 × 10³ U g⁻¹. The use of a CaCl₂ solution yielded 5.99 × 10³ U g⁻¹ of membrane activity, which showed that the immobilized lipase needed some ionic strength to maintain its activity. It was because ionic strength in the aqueous phase can help the enzyme keep its catalytic conformation. PBS (0.05 M, pH 7.0) was found to be the best providing a membrane activity as high as 1.46 × 10⁴ U g⁻¹. These results indicate that the immobilized lipase prefers a solution with a certain ionic strength as well as a relatively stable pH in order to maintain its activity. Moreover, a titration of the PBS (0.05 M, pH 7.0) with the standard NaOH solution showed a gradual increase in the aqueous pH without sudden changes, indicating the feasibility of this system for extracting and neutralizing the produced fatty acid (Fig. S3†). As shown in Table 1, the three R values were all >0.98, demonstrating a good accord with the pseudo first order model.

Fig. 4 Effect of operation conditions on bioreactor efficiency measured by fatty acid production versus time. (a) Aqueous-phase system (pure water, CaCl2 solution and PBS pH 7.0). (b) Aqueous phase (pH 6.0, pH 7.0 and pH 8.0). (c) Temperature (11 °C, 25 °C, 40 °C and 55 °C). (d) Membrane amount (4, 11, 21, 30, 58 and 77 mg).

Table 1 Effect of operation conditions on bioreactor efficiency

Operation condition		q1 (mmol g^−1 or mmol)	k1 (min^−1)	R
Aqueous-phase system	Pure water	5.97 × 10^4	3.94 × 10^−7	0.988
	CaCl2 solution	21.0	1.28 × 10^−2	0.998
	PBS7.0	48.7	1.76 × 10^−2	0.993
Aqueous phase pH	PBS6.0	32.6	1.64 × 10^−2	0.979
	PBS7.0	48.7	1.76 × 10^−2	0.993
	PBS8.0	48.8	1.10 × 10^−2	0.995
Temperature	11 °C	30.3	1.64 × 10^−2	0.993
	25 °C	48.7	1.76 × 10^−2	0.993
	40 °C	54.1	2.56 × 10^−2	0.968
	55 °C	48.1	3.57 × 10^−2	0.980
Weight of lipase-immobilized membrane	4 mg	9.28 × 10^2	9.1 × 10^−3	0.990
	11 mg	7.63 × 10^2	1.24 × 10^−2	0.985
	21 mg	8.18 × 10^2	1.94 × 10^−2	0.991
	30 mg	7.70 × 10^2	1.50 × 10^−2	0.993
	58 mg	3.03 × 10^2	1.10 × 10^−2	0.978
	77 mg	2.81 × 10^2	2.62 × 10^−2	0.990

---

3.4 Effect of aqueous phase pH on bioreactor efficiency

The pH is one important parameter capable of altering enzymatic activities in aqueous solution. The fatty acid amount produced by the lipase-immobilized membrane as a function of pH is depicted in Fig. 4(b). PBS (0.05 M, pH 6.0), PBS (0.05 M, pH 7.0) and PBS (0.05 M, pH 8.0) were tested as the aqueous solutions, yielding corresponding membrane activity of 1.05 × 10⁴ U g⁻¹, 1.46 × 10⁴ U g⁻¹ and 1.36 × 10⁴ U g⁻¹. The instantaneous neutralization of fatty acid by NaOH at the aqueous solution site helped to accelerate the extraction of hydrolytic products on the aqueous side. Moreover, the increase of OH⁻ concentration should enhance the driving force for the diffusion of OH⁻ from the bulk to the reaction front, the transformation of fatty acid RCOOH to RCOO⁻ at the reaction site and the diffusion of RCOO⁻ from the reaction front to the aqueous bulk phase.¹⁰ It is likely that all of these factors contributed to the high catalytic efficiency at a higher pH. Nevertheless, when the pH increased above a threshold, immobilized lipases on the membrane tended to be unstable and this unfavorably affected the hydrolysis conversion resulting in a drop in membrane activity at pH 8.0 compared to pH 7.0. The fitting parameters for this process are depicted in Table 1, with R values of 0.979, 0.993 and 0.995 as the pH increased from 6 to 8.

3.5 Effect of temperature on bioreactor efficiency

Fig. 4(c) shows the fatty acid production as a function of time and temperature. The optimal activity temperature of the free lipase was 30 °C; and according to Fig. 4(c), we can see the immobilization process can greatly improve the thermal stability of the lipases. It was observed the activity of lipase-immobilized membrane increased when the temperature increased from 11 °C to 40 °C, then decreased when the temperature went higher.

Compared with the membrane activity of 9.26 × 10³ U g⁻¹ at 11 °C, there was an almost two times enhancement to 1.85 × 10⁴ U g⁻¹ at 40 °C. The high temperature resistance of the membrane may be explained by the fact that the presence of covalent bonds between the enzyme and the nanofiber membrane reduced the flexibility of the enzyme conformations and thus a higher temperature was needed for the movement of the lipase molecule to assume a catalytically active conformation.¹³ Another possible explanation is that the transfer resistances for the substrate and hydrolysis product were diminished at a higher temperature, which can improve the bioreactor efficiency.¹⁷ However, a further increase of the temperature reduced the stability of the lipase and resulted in its conformational denaturation, thus decreasing the lipase activity. In fact the membrane activity decreased to 1.58 × 10⁴ U g⁻¹ at 55 °C (the fitting curves are also shown in Table 1).

3.6 Effect of inter-fiber transfer limitation on bioreactor efficiency

We considered the inter-fiber transfer limitation as mass transfer of the substrate to and product from the immobilized enzyme may also affect the bioreactor efficiency. In this immobilization system, inter-fiber transfer limitation can be facilely studied by varying the electrospinning time to control the nanofiber membrane thickness. Membranes with weights of 4 mg, 11 mg, 21 mg, 30 mg, 58 mg and 77 mg were installed and their fatty acid productions were measured (the membrane thickness was 13.9 μm, 47.4 μm, 85.2 μm, 114 μm, 226 μm and 300 μm, respectively). The tested membranes were all electrospun under the same conditions, although for different electrospinning times, thus they shared the same fiber diameter and porosity. Results are shown in Fig. 4(d). When the membrane amount was increased, more activated acrylic acid groups were available to couple with the lipase molecules, thus it was speculated that a relatively higher hydrolysis conversion could be obtained. However, as shown in the figure, the bioreactor efficiency was only slightly affected when the membrane weight was between 4 and 30 mg, demonstrating no significant transfer limitation. Further study showed that the pores of the nanofiber membrane were too small for olive oil droplets to permeate, so the hydrolysis reaction only took place on the membrane surface and changes in membrane weight did not yield any variations in bioreactor efficiency. When the membrane weight increased to 58 mg, the transfer resistance for the hydrolysis products got apparent and less fatty acid was extracted, moreover, the least product was extracted when the membrane weight was 77 mg (the fitting parameters are shown in Table 1).

From the observations detailed above, it can be inferred that the efficiency of this biphasic bioreactor can be improved by increasing the contact area of the membrane with the substrate emulsion as a way to increase the amount of membrane actually in use. Also the membrane weight should be carefully chosen to reach a balance between sufficient physical strength for bioreactor operation and high hydrolysis conversion.

3.7 Explanation of the immobilized lipase performance

Although most kinetic models are based on simplified systems containing only a few different chemical species (which is because the enormous amounts of experimental data and numerical work necessary to conduct the associated regression analysis would be prohibitive and difficult to handle), the pseudo first order model applied in this work was suitable as an aid for analysis of the immobilized lipase.

Previous work has reported that free lipase from Candida rugosa obeys first order kinetics for the hydrolysis of tallow, coconut and olive oil.^23,24 In this work, the immobilized lipase followed a similar model, with most R values >0.98. Additionally, a report indicated that a pseudo first order model was not suitable to model the kinetics of lipase at temperatures higher than 50 °C due to the deactivation of free lipase.²⁵ In this work, the immobilized lipase still performed well and followed the first order model at 55 °C. Similarities between the catalytic mechanisms of free and immobilized lipases and the good hydrolysis performance of the immobilized lipases in this work may be ascribed to the characteristics of the nanofiber membranes. During the enzyme immobilization process, the support plays a role in the creation of an enzyme-rich phase. When a support is used to immobilize an enzyme like a lipase, the three levels of structure that must be considered are the macroscopic level, microscopic level, and submicroscopic level.²⁶ At the macroscopic level, the PANCAA nanofiber membrane has two dominant dimensions. In terms of microscopic characteristics, two factors are important to solid supports: the thickness and the porous structure.²⁶ Nanofiber membranes are usually endowed with minimal thickness, high specific surface area and porosity, which can offer accessibility to the active site, minimize substrate diffusion resistance and maximize the available area for lipase attachment.^14,15 Moreover, compared with other porous materials such as mesoporous activated carbon, which has also been used for lipase immobilization, nanofiber membranes have the benefit of following a pseudo first order kinetic model while providing good product separation and catalytic system recycling.

3.8 Time course of olive oil hydrolysis in bioreactor

One problem with the hydrolysis reaction is the fast decrease in the reaction rate as the conversion of oil to fatty acid increases.²⁷ In order to test whether the lipase-immobilized membrane in the bioreactor can continuously catalyze the conversion, a time course study was carried out for various time intervals ranging from 1 h to 19 h. The membrane was 20 mg and the reaction was performed at 25 °C using PBS (0.05 M, pH 7.0) as the aqueous solution. The result is shown in Fig. 5. The production of fatty acid from olive oil increased during the test time. The rate of fatty acid production declined after 2 h and was stable after 8 h of reaction. Upon examining the entire duration of the reaction, we found that the hydrolysis reaction rate did not decrease significantly as the reaction progressed. These results verified the highly stable and efficient nature of the lipase-immobilized membrane system for the hydrolysis of olive oil and indicated the great potential of this system for catalyzing a wide range of reactions.

Fig. 5 Time course of olive oil hydrolysis in the bioreactor.

3.9 Reusability of the lipase-immobilized nanofiber membrane in the bioreactor

The ability to reuse the immobilized lipase is important for ensuring that it is economical to employ the enzyme in repeated batch or continuous reactions.¹¹ If the immobilized lipase has a relatively long lifetime, the cost will be significantly decreased and its industrial implementation will be accelerated. In the reusability studies, the membrane activity in the first reaction was set as 100%, and the hydrolysis conversions in the subsequent reactions were compared to it. A slight decrease in membrane activity appeared after the second usage and after 6 cycles of batch operation, 88% of the original activity of the lipase-immobilized membrane remained (Fig. 6). It is possible that this activity loss is related to the inactivation of the lipase by continuous use and loss of the membrane fabric.¹⁷ Activity retentions of 50% after three reuses for the catalysis of olive oil hydrolysis,¹⁰ 80% after five reuses for fish oil hydrolysis²⁸ and 52% after six reuses for vegetable oil hydrolysis²⁹ have been reported for lipases immobilized on various supports. In comparison, the immobilized lipase reported in this work provided a significant advantage in stability for batch recycling.

Fig. 6 Reusability of the lipase-immobilized nanofiber membrane in the bioreactor.

4. Conclusions

The current study comprehensively investigated the properties of a biphasic lipase-immobilized PANCAA nanofiber membrane bioreactor using the hydrolysis of olive oil as a model reaction. Membrane diameters and operation parameters were optimized, achieving a bioreactor activity of 1.85 × 10⁴ U g⁻¹. The bioreactor showed good operation stability and 88% of the original membrane activity after 6 batch reactions. The results indicated that a certain balance exists when attempting to optimize both membrane and operation parameters. This work should prove useful in providing information on the underlying principles associated with maximizing the utility of immobilized enzymes in a bioreactor for industrial scale applications.

Acknowledgements

The authors are grateful to the financial support from the Fundamental Research Funds for the Central Universities (Grant no. 2013QNA4090), the National Natural Science Foundation of China (Grant no. 21274126), and the National “Twelfth Five-Year” Plan for Science & Technology Support of China (Grant no. 2012BAI08B01).

Notes and references

1. S. J. Benkovic and S. Hammes-Schiffer, Science, 2003, 301, 1196.
2. A. Schmid, J. S. Dordick, B. Hauer, A. Kiener, M. Wubbolts and B. Witholt, Nature, 2001, 409, 258.
3. X. J. Huang, A. G. Yu and Z. K. Xu, Bioresour. Technol., 2008, 99, 5459.
4. L. Ragupathy, B. Pluhar, U. Ziener, H. Keller, R. Dyllick-Brenzinger and K. Landfester, J. Mol. Catal. B: Enzym., 2010, 62, 270.
5. R. Fernandez-Lafuente, C. M. Rosell, V. Rodriguez, C. Santana, G. Soler, A. Bastida and J. M. Guisán, Enzyme Microb. Technol., 1993, 15, 546.
6. J. M. Guisán, Enzyme Microb. Technol., 1988, 10, 375.
7. X. J. Huang, P. C. Chen, F. Huang, Y. Ou, M. R. Chen and Z. K. Xu, J. Mol. Catal. B: Enzym., 2011, 70, 95.
8. A. Kumar and S. S. Kanwar, Bioresour. Technol., 2011, 102, 2162.
9. X. J. Huang, Z. K. Xu, X. D. Huang, Z. G. Wang and K. Yao, Polymer, 2006, 47, 3141.
10. P. C. Chen, X. J. Huang, F. Huang, Y. Ou, M. R. Chen and Z. K. Xu, Cellulose, 2011, 18, 1563.
11. T. F. Jiao, B. D. Leca-Bouvier, P. Boullanger, L. J. Blum and A. P. Girard-Egrot, Colloids Surf., A, 2010, 254, 284.
12. C. C. Lee, H. P. Chiang, K. L. Li, F. H. Ko, C. Y. Su and Y. S. Yang, Anal. Chem., 2009, 81, 2737.
13. A. F. Wait, A. Parkin, G. M. Morley, L. dos Santos and F. A. Armstrong, J. Phys. Chem. C, 2010, 114, 12003.
14. S. Agarwal, A. Greiner and J. H. Wendroff, Adv. Funct. Mater., 2009, 19, 7198.
15. N. Bhardwaj and S. C. Kundu, Biotechnol. Adv., 2010, 28, 1315.
16. Z. G. Wang, L. S. Wan, Z. M. Liu, X. J. Huang and Z. K. Xu, J. Mol. Catal. B: Enzym., 2009, 56, 189.
17. P. Jochems, Y. Satyawali, L. Diels and W. Dejonghe, Green Chem., 2011, 13, 1609.
18. M. Bradford, Anal. Biochem., 1976, 72, 248.
19. Y. S. Ho and G. McKay, Process Saf. Environ. Prot., 1998, 76, 332.
20. K. Ramani, S. Karthikeyan, R. Boopathy, L. J. Kennedy, A. B. Mandal and G. Sekaran, Process Biochem., 2012, 47, 435.
21. J. L. Liu, J. Q. Wang, L. G. Bachas and D. Bhattacharyya, Biotechnol. Prog., 2001, 17, 866.
22. Y. H. Wang, J. C. Zhang and J. M. Yin, Desalin. Water Treat., 2009, 1, 157.
23. W. M. Linfield, R. A. Barauskas, L. Sivieri, S. Serota and R. W. Stevenson, J. Am. Oil Chem. Soc., 1984, 61, 191.
24. W. M. Linfield, D. J. Obrien, S. Serota and R. A. Barauskas, J. Am. Oil Chem. Soc., 1984, 61, 1067.
25. L. Dandik and H. A. Aksoy, J. Am. Oil Chem. Soc., 1992, 69, 1239.
26. V. M. Balcão, A. L. Paiva and F. X. Malcata, Enzyme Microb. Technol., 1996, 18, 392.
27. M. S. Puthli, V. K. Rathod and A. B. Pandit, Biochem. Eng. J., 2006, 31, 31.
28. J. Y. Yan, S. X. Liu, J. Hu, X. H. Gui, G. L. Wang and Y. J. Yan, Bioresour. Technol., 2011, 102, 7154.
29. M. Yiğitoğlu and Z. Temoçin, J. Mol. Catal. B: Enzym., 2010, 66, 130.

---

Footnote

† Electronic supplementary information (ESI) available: Scanning electron microscope (SEM) micrographs of the nanofiber membranes with different fiber diameters, SEM micrographs of the nanofiber membranes during reactions in the bioreactor, and titration of PBS (0.05 M, pH 7.0) with NaOH. See DOI: 10.1039/c3ra46779a

---