Evaluating effective factors on the activity and loading of immobilized α-amylase onto magnetic nanoparticles using a response surface-desirability approach

Abstract

The effects of different operational conditions of α-amylase covalent immobilization on magnetic nanoparticles (MNPs), such as initial enzyme concentration, glutaraldehyde (GA) concentration, pH, and ionic strength were investigated using a central composite design (CCD). Moreover, two responses, the biocatalyst activity and amount of immobilized enzyme were simultaneously studied using Derringer’s desirability function. The optimum amount and activity of immobilized enzyme were determined to be 24.83% and 556.41 mg g_MNP⁻¹ at an initial enzyme concentration of 999.86 ppm, solution pH of 4.6, GA concentration of 0.59%, ionic strength of 99.99 mM and a process time of 4 h. The study of the kinetic parameters and enzyme stability showed a significant enhancement in the performance of the immobilized enzyme with respect to the free enzyme. The storage stability and reusability of the immobilized biocatalyst were found to be about 50 and 40% of the initial activity after 12 days and 6 cycles, respectively.

---

1. Introduction

Enzyme immobilization is one of the advantageous methods that can improve the operation of biocatalysts in industrial bioprocesses. It facilitates enzyme separation from the reaction medium for recycling, stabilization in extreme conditions such as high pH or temperature, the protection of enzymes against denaturing agents that can destroy the active site, and increases lifetime and thereby reduces the operating costs.¹

Amylases are one of the most significant hydrolase enzymes, occupying nearly 30% of the global enzyme market.² α-Amylase (EC 3.2.1.1) is one of the amylase family which catalyzes the hydrolysis of internal α-1,4-glycosidic bonds in starch molecules and similar carbohydrates to maltose, glucose and other low molecular weight products.^3–5 α-Amylase has a variety of applications such as starch saccharification, in food, baking, fermentation, detergents, textiles and the paper industry.⁶ Owing to the wide industrial applications, immobilization of α-amylase is a vital step to overcome the problems of using the free enzyme. The properties of immobilized biocatalysts depend considerably on the immobilization method and its conditions. There are many reports on the covalent immobilization of α-amylase on different supports, illustrating the advantages of the immobilized enzyme such as increased rigidity, retaining enzyme activity, and prevention of leakage and unfolding during the catalytic process.^7,15

It should be noted that in previous studies, a limited number of factors such as enzyme concentration, time and pH were investigated for the optimization of α-amylase^2,4,7,8 So, despite the various research that has been conducted on covalent α-amylase immobilization, there is a vacancy for a detailed investigation on different immobilization conditions and their interaction effects on the resultant biocatalyst efficiency. Moreover, optimization of different operating conditions with an analysis of the interactions between the factors affecting the activity and the amount of immobilized enzyme simultaneously could be very important for the development of an optimized biocatalyst.^3,4,7 In our recent work, the effects of various immobilization conditions on covalent bond formation were investigated for α-amylase immobilization on magnetic nanoparticles (MNP).⁷ However, it is vital to determine the optimum immobilization conditions leading to a perfect biocatalyst.

Response surface methodology (RSM) is recommended for multivariate studies because of its ability to produce empirical models and analyse the response to problems including several process factors using the approach of response optimization. It is a collection of mathematical and statistical techniques which has been widely applied to optimize and evaluate the interactive effects of independent variables in numerous enzymatic processes.^6,9–11

The multi-response surface method is used for solving the optimization problems of several responses. This methodology is applied when various responses have to be considered at the same time and there is a necessity for finding optimal compromises between the total numbers of responses taken into account.¹²

In this study, a multi-criteria decision making approach, Derringer’s desirability function, was used for the evaluation of two different responses (the amount of immobilized enzyme and its activity). The applications and advantages of the desirability functions have already been discussed in different informative articles.¹³ To the best knowledge of the authors, there are no works on optimizing the amount of enzyme loading and activity simultaneously with the RSM and desirability functions.

For this, the covalent immobilization conditions of α-amylase on the magnetic nanoparticles were studied in detail using a central composite design (CCD) under response surface methodology (RSM). Aminated magnetic nanoparticles activated by glutaraldehyde (GA) were used as a support for immobilization. The effect of different factors (initial enzyme concentration, GA concentration, pH, ionic strength and immobilization time) and their interactions on the amount and activity of the immobilized enzyme were investigated simultaneously using the statistical approach of RSM, and multi-response optimization was done using Derringer’s desirability function. Some significant interactions between the main factors on the two responses were evaluated. Also, the different stabilities of α-amylase, the reusability of the biocatalyst and the kinetic factors of the immobilized and free enzyme were studied under the optimum conditions.

2. Materials and methods

2.1. Materials

All chemicals, buffers and other reagents were of analytical grade, and were used without further purification. Ferrous chloride (FeCl₂·4H₂O) and ferric chloride (FeCl₃·6H₂O) were used as an iron source. Aqueous ammonia (25% (v/v) aqueous solution) and ethanol were used as a precipitation agent and solvent, respectively. All of these chemicals were purchased from Merck. Crude α-amylase from Aspergillus oryzae (EC 3.2.1.15, activity 35.7 U mg⁻¹), GA, maltose, starch and 3,5-dinitro-salicylic acid (DNS) were purchased from Sigma-Aldrich. Deionized water (DI-water) was used in the preparation of all solutions.

2.2. MNP synthesis and coating

The synthesis procedure and surface treatment of the MNPs were reported in our previous work.⁷ First, the MNPs were synthesized using the co-precipitation method and were covered with amino-silane functional groups via the silanization reaction using 3-aminopropyltriethoxysilane (APTES). After that, aminated MNPs were activated with different amounts of GA (0.5 and 10%) in 20 mM phosphate–citrate buffer (pH 7) according to the design of the experiments.

2.3. Enzyme immobilization

Two milliliters of enzyme solution (250–1000 ppm) was added to 2 mg of MNPs as a synthesized support which was equilibrated with 5–100 mM of phosphate–citrate buffer (pH 4.6–7.6). Thereafter, the mixture was shaken with a rotation speed of 200 rpm at 30 °C for different time periods (1–4 h) according to the design of the experiments. All the experiments were done in triplicate and the mean values are reported as the results. The amount of immobilized α-amylase on the support was determined using Bradford’s method.¹⁴

2.4. Analysis methods and MNP characterization

2.4.1. Amylase activity assay. The activities of free and immobilized α-amylase were determined using the method proposed by Miller.¹⁵ The number of reduced maltose molecules which were produced in the hydrolysis reaction of starch were measured using 3,5-dinitrosalicylic acid (DNS). The amount of produced maltose was determined using UV spectroscopy at a wavelength of 540 nm. Amylase activity is defined as micromoles of maltose produced within 1 min at 30 °C.

2.4.2. Characterization of MNPs. The surface chemical groups were identified with Fourier transform infrared spectroscopy (FT-IR). The FT-IR spectra were recorded using a Nicolet NEXUS 670 instrument. A Philips PW1800 diffractometer with Cu-Kα radiation (0.154056 nm) was used for the X-ray diffraction (XRD) of nanoparticles. The morphology and particle size distribution of the support material and biocatalyst were determined using transmission electron microscopy (TEM) and field emission scanning electron microscopy (FESEM). TEM imaging of the MNPs was carried out with a Philips-CM30 microscope and FESEM micrographs were recorded using a Philips-XL20.

2.5. Experimental design and statistical analysis

Preliminary experiments were carried out to determine the factors affecting the amylase immobilization process (data not presented here). A full factorial design was carried out to evaluate the effects of five important variables in two levels including initial enzyme concentration (250–1000 ppm), GA concentration (0.5–10%), pH (4.6–7.6), ionic strength (5–100 mM) and immobilization time (1–4 h) on two responses of the amount of immobilized enzyme and its activity. The experimental design and analysis was done using ‘Design Expert’ software (version 8.0). 36 experiments were designed including 32 trials of full factorial designs (2⁵) and four replicates at the central points for the full factorial model. All tests were done in duplicate and the analysis of variance was performed based on these duplicate results. After ensuring that the curvature is important in the results of the full factorial design, the levels of the immobilization factors and the interaction effects were analyzed and optimized using the central composite design (CCD) of the response surface methodology (RSM). For this purpose, 10 trials of axial points were added to the full factorial experimental design for converting it to RSM with α = 1 (Table 1). A series of experiments at given optimal conditions were performed in order to study the validation of the response surface model and compare the predicted values and experimental data.

Table 1 CCD of experiments and the amount and activity of immobilized α-amylase

Trial	Cenz (ppm)	CGA (%)	Time (h)	pH	Cbuffer (mM)	Immob. (mg gMNP^−1)	Act. (%)
1	250	0.5	1	4.6	5	76.46	0.50
2	1000	0.5	1	4.6	5	517.18	7.85
3	250	10	1	4.6	5	82.15	12.07
4	1000	10	1	4.6	5	591.51	11.86
5	250	0.5	4	4.6	5	75.09	10.69
6	1000	0.5	4	4.6	5	551.55	16.53
7	250	10	4	4.6	5	140.79	12.48
8	1000	10	4	4.6	5	667.34	17.89
9	250	0.5	1	7.6	5	81.96	0.43
10	1000	0.5	1	7.6	5	511.68	6.67
11	250	10	1	7.6	5	13.39	11.99
12	1000	10	1	7.6	5	466.13	13.80
13	250	0.5	4	7.6	5	88.83	11.79
14	1000	0.5	4	7.6	5	540.55	15.27
15	250	10	4	7.6	5	104.39	13.52
16	1000	10	4	7.6	5	593.53	18.17
17	250	0.5	1	4.6	100	101.20	1.50
18	1000	0.5	1	4.6	100	536.43	18.58
19	250	10	1	4.6	100	73.05	10.75
20	1000	10	1	4.6	100	607.68	12.07
21	250	0.5	4	4.6	100	89.52	8.88
22	1000	0.5	4	4.6	100	551.55	13.29
23	250	10	4	4.6	100	123.61	13.66
24	1000	10	4	4.6	100	656.22	16.93
25	250	0.5	1	7.6	100	79.21	0.00
26	1000	0.5	1	7.6	100	517.18	8.48
27	250	10	1	7.6	100	20.48	11.37
28	1000	10	1	7.6	100	465.12	10.89
29	250	0.5	4	7.6	100	86.21	5.80
30	1000	0.5	4	7.6	100	533.68	23.10
31	250	10	4	7.6	100	77.09	11.51
32	1000	10	4	7.6	100	523.76	17.27
33	250	5.25	2.5	6.1	52.5	134.48	7.46
34	1000	5.25	2.5	6.1	52.5	574.82	9.11
35	625	0.5	2.5	6.1	52.5	240.52	9.71
36	625	10	2.5	6.1	52.5	242.9	9.08
37	625	5.25	1	6.1	52.5	281.21	8.69
38	625	5.25	4	6.1	52.5	340.8	6.47
39	625	5.25	2.5	4.6	52.5	340.172	7.74
40	625	5.25	2.5	7.6	52.5	291.55	10.42
41	625	5.25	2.5	6.1	5	373.62	8.09
42	625	5.25	2.5	6.1	100	389.13	9.47
43	625	5.25	2.5	6.1	52.5	308.52	6.38
44	625	5.25	2.5	6.1	52.5	301.44	7.14
45	625	5.25	2.5	6.1	52.5	312.56	8.18
46	625	5.25	2.5	6.1	52.5	321.66	6.45

---

2.6. Stability tests

2.6.1. Thermal and pH stability. The thermal stability of the enzyme immobilized using the optimum conditions and the free enzyme was studied in the temperature range of 30–80 °C. For this purpose, their residual activities of the enzymes were measured after 30 min incubation in phosphate–citrate buffer (20 mM, pH = 6) at each temperature.

The pH stability was investigated by evaluating the residual activity of the enzyme after 30 min of incubation at a certain pH in the range of 4–8 and at 30 °C.

2.6.2. Storage stability and reusability. The storage stability of the free and optimally immobilized α-amylase was investigated by measuring the residual activity (calculated as a percentage of the initial activity) at different times (1–12 days). In addition, the reusability of the immobilized enzyme was evaluated using an activity test of six cycles. The immobilized α-amylase was added to 2 ml of 1% (w/w) starch in 20 mM phosphate–citrate buffer (pH = 6) and incubated for 5 min with a constant shaking speed at 30 °C. The amount of the immobilized enzyme released during the activity test was measured in each cycle via the Bradford method. The DNS method was used for measuring the relative activity of the residual immobilized enzyme in each cycle. After each cycle, the immobilized enzyme was settled using a strong magnet, washed with distilled water and then used for the next cycle.

2.7. Kinetic parameters

The activity of the optimally immobilized and free enzyme was measured in different substrate concentrations (0.1–0.5 mg ml⁻¹). The K_m and V_max values were calculated from a Lineweaver and Burk plot using the Michaelis–Menten model:where [S] was the concentration of the substrate and K_m was the Michaelis constant. V and V_max represented the initial and maximum rate of reaction, respectively.

3. Results and discussion

3.1. MNP characterization

The FTIR spectra of aminated MNPs (a) activated by GA (b) and immobilized α-amylase (c) are shown in Fig. 1. The peak observed at 1630 cm⁻¹ corresponds to the bending mode of NH₂. The peak observed at 3438 cm⁻¹ confirms the presence of N–H bonds.^16,17 The Fe–O bonding of activated magnetic nanoparticles is indicated by two overlapped peaks at 590 and 610 cm⁻¹.¹⁷ These peaks disappear after enzyme immobilization (Fig. 1c), confirming the full surface coverage of the activated MNPs by amylase. Also, the broad peak around 1112 cm⁻¹ indicates surface coverage by α-amylase after immobilization. The absence of the characteristic peaks of GA and MNPs in Fig. 1c suggests the covalent bonding of GA and amylase.⁷

Fig. 1 FT-IR of bare MNP (a), GA activated MNP (b) and immobilized α-amylase on MNP (c).

Fig. 2 shows the XRD pattern of modified MNPs with the corresponding diffraction planes. The characteristic diffraction peaks are assigned according to the reference pattern of JCPDS 01-087-0246. The crystalline structure of the as-prepared magnetic nanoparticles is cubic. In addition, the sharp diffraction peaks confirm the perfect crystallinity of the particles with an average crystallite size of 11.8 nm calculated using the Scherrer equation. Perfect matching of the diffraction peaks with the reference pattern demonstrates the phase purity of the MNPs after surface modification.

Fig. 2 X-ray pattern of modified MNPs.

Fig. 3 shows the TEM micrograph and particle size distribution histogram of the MNPs before and after immobilization. It demonstrates the narrow size distribution and uniform shape of the MNPs. As shown in Fig. 3a, before enzyme immobilization most of the particles have a size in the range of 12 to 20 nm; the particle size distribution graph has a maximum at 15 nm. After enzyme loading on the MNPs, a particle size increase was observed; as shown in Fig. 3b, the maximum of the particle size histogram was shifted from 15 to 20 nm. This particle size increase could be attributed to linker overlay and also to the immobilized enzyme on the MNPs.

Fig. 3 TEM images and particle size distribution histograms of MNPs before (a) and after (b) immobilization.

Fig. 4 shows the SEM images of the modified MNPs before and after enzyme immobilization. Fig. 4a shows uniform spherical particles; however, the SEM images in Fig. 4 demonstrate particle agglomeration with an average size of 50 nm compared to Fig. 3, which could be due to the presence of surface hydroxyl and amine groups, as demonstrated by the FT-IR results. After immobilization (Fig. 4b), the shape of the amylase covered MNPs remained constant, whereas the average size of the agglomerated particles increased to 70 nm.

Fig. 4 SEM images of MNPs before (a) and after (b) immobilization.

3.2. Statistical analysis of experimental designs

Studying all of the effective factors and their interactions is necessary to optimize the immobilization process. The factors include the initial enzyme concentration (C_enz), GA concentration (C_GA), pH, ionic strength (C_buffer) and time.

Two steps of the experiment were designed to facilitate the optimization of the α-amylase immobilization process. At first, the full factorial design with 32 experiments (the first 32 experiments in Table 1) and 4 experiments at the center point levels (the last four experiments in Table 1) was used to study the effects of five factors on the amount and activity of the immobilized enzyme.

After confirming that the curvature is important in the ANOVA results of the immobilized α-amylase activity in the full factorial design (results not presented here), 10 experiments related to the axial points of the CCD were added to the full factorial experimental design (experiment number 33–42 in Table 1). The levels of the immobilization factors and the results representing the amount and activity of the immobilized enzyme were obtained (Table 1) and the main and interaction effects were analyzed and optimized (Tables 1-S and 2-S in the ESI†) based on the central composite design (CCD).

A second order mathematical equation as a result of the combination of 46 tests including the regression coefficient (β) was obtained via the least square method. The model was evaluated using ANOVA for the amount of immobilized enzyme and its activity, and the results are shown in Table 1-S and 2-S† respectively.

The variance analysis of the quadratic regression model and low probability value of Fischer’s F-test (<0.0001) indicate the high significance of the applied model. The F-values of 29.06 and 545.91 demonstrated that the independent factors and their interactions had a significant effect on the activity and the amount of the immobilized enzyme. The fitness of the model was investigated using the determination coefficient (R-squared). In this case, the value of the determination coefficients, 0.8713 for the activity and 0.9958 for the amount of immobilized enzyme, indicate that 12 and 1% of the total coefficients could not be determined by this model, respectively. In addition to this, the high values of the adjusted determination coefficients (Adj R-squared) represent the great significance of the model. The “Pre R-squared” was close to the “Adj R-squared”.

The F-values of “Lack of Fit” of 1.56 and 0.0736 for the activity and amount of the immobilized enzyme, respectively, suggest less significance relative to pure error. In addition to the lack of fit tests, the model was further evaluated using the observed vs. predicted plot. The points of all of the predicted and actual responses fell in 45° lines, also indicating good agreement with the model (results not presented here).

3.3. Effect of various factors on enzyme immobilization

3.3.1. Enzyme and GA concentrations. According to the main effects plots (not presented), the enzyme and GA concentrations have positive effects on the activity of the immobilized enzyme. Fig. 5(a and b) shows the 3D surface plot of the GA and enzyme concentration effects on the amount of immobilized enzyme and its activity. As shown in Fig. 5(a and b), the amount and activity of the immobilized enzyme increased with the initial enzyme concentration in all GA concentration levels, but the activity increase rate is higher in the case of low concentrations of GA in comparison to high GA concentration.

Fig. 5 Response surface plots representing the interaction of enzyme and GA concentration (a and b), enzyme concentration and time (c and d), GA concentration and time (e and f), enzyme concentration and ionic strength (g), and enzyme concentration and pH (h) on the activity and amount of the immobilized enzyme respectively.

In other words, a higher level of GA concentration results in a lower activity when the initial enzyme concentration increases (Fig. 5a). Fig. 5b demonstrates that immobilized enzyme loading increases with the GA concentration until the center point, then decreases.

The increase of proper linkage sites via GA concentration increase results in activity and enzyme immobilization increase. Then the concentrating of GA on the surface of MNPs leads to the formation of GA dimer molecules and a decrease of linkage site availability.¹⁸ This phenomenon results in lower enzyme loading on the MNPs, however a higher mobility of the enzyme bonded to GA leads to higher activity.^7,19 Also, a low concentration of enzyme increases the possibility of single-point covalent bond formation leading to lower enzyme denaturation.¹⁹

In high enzyme concentration, GA concentration increase decreases the activity of biocatalyst. According to Nwagu and coworkers,⁸ crosslinking between the enzymes via GA molecules leads to a higher rigidity of the enzyme and hence, lower activity. In addition, the high concentration of GA could change the globular structure of the enzyme during the immobilization process and therefore change its activity.²⁰

3.3.2. Enzyme concentration and time. Fig. 5(c and d) shows the effect of immobilization time on enzyme activity and loading. In different initial enzyme concentrations, the immobilized enzyme activity increases with the immobilization time increase (Fig. 5c). Fig. 5d illustrates that the immobilization time has no significant impact on the amount of immobilized enzyme.

It can be concluded from Fig. 5c that biocatalyst activity with higher process times is more affected by the enzyme concentration in comparison to lower times. It has been reported that low concentrations of enzyme during the immobilization process increase the possibility of bonding between the enzyme active site and a linker leading to the lower activity of the biocatalyst. Moreover, it is possible to form multiple bonds between GA molecules and one enzyme. This means a decrease in the activity of the enzyme through increased rigidity.⁸ In addition, increasing immobilization process time could increase the possibility of active site blockage.

Fig. 5(c and d) demonstrates that increasing enzyme loading on the support does not increase the biocatalyst activity in the same way. This shows that immobilization time increase simultaneously increases enzyme loading and active site blockage. Therefore, activity enhancement is not as significant as enzyme loading.

3.3.3. GA concentration and time. According to Fig. 5e, in a short-time experiment, the immobilized biocatalyst activity increases with GA concentration. But in a long-time experiment (4 hours), the activity decreases. In both levels of time, the amount of immobilized enzyme increases at first with an increasing GA concentration, and then decreases (Fig. 5f). This behavior corresponds to the immobilization mechanism change from ionic adsorption to covalent bonding by the GA concentration increase.⁷

The ascending activity rate of the biocatalyst due to GA concentration in short immobilization times can be attributed to enzyme loading. Fig. 5f shows a maximum enzyme loading around the central points. Despite enzyme loading decreasing after the central point, the activity increase of the biocatalyst was observed (Fig. 5e). This inconsistency could correspond to linker elongation via the increasing GA concentration. This leads to the higher mobility and reduced steric hindrance of the anchored enzyme.^7,19

The biocatalyst activity reaches a minimum around the central point. The falling rate can be attributed to a bonding increase between the linker and the enzyme, leading to higher rigidity. Decreasing the enzyme loading enhances the biocatalyst activity via mobility enhancement.^8,21

3.3.4. pH and ionic strength. According to ANOVA (Tables 1-S and 2-S†), the pH and ionic strength of the immobilization medium were not effective factors for biocatalyst activity and enzyme loading, respectively. The interaction plots are shown in Fig. 5(g and h). As shown in Fig. 5g, while the initial concentration of the enzyme is low, high ionic strength solutions result in a lower activity in comparison to low ionic strength solutions. The immobilizing mechanism of the enzyme on the GA activated support affects the activity of the biocatalyst; since the ionic strength of solution is one of the important determining factors on the bonding type, it is an effective factor in the activity.²²

Physical bond formation is more favorable than covalent bonding in low ionic strength solutions and with a high number of reactive groups; so, in low ionic strength solutions immobilization is initiated via physical adsorption and then continues with covalent bond formation. But in high ionic strength solutions, covalent bonding is the primary mechanism of immobilization.⁷ So, when the initial concentration of the enzyme is low, high ionic strengths result in covalent bond formation; therefore in this case the biocatalyst shows a lower activity in comparison to the low ionic strength conditions. But in high enzyme concentrations, the high number of reactive amino acid groups leads to more physical bonds forming and results in higher activity.²²

Fig. 5h shows that low pH of the immobilization medium leads to a higher enzyme loading. This behavior can be attributed to the surface charge of the support and enzyme. According to the manufacturer’s data sheet, the isoelectric point of α-amylase is 5.4. This indicates that the surface charge of the enzyme in high level pH is negative and in low level pH is positive. Also, a negative surface charge of the activated support was observed in both high and low level pH via zeta potential measurement.⁷ The surface charge data predicts that low pH immobilization media increases the enzyme loading via ion–ion interactions, and also a higher pH than 5.4 will retard the immobilization rate via electrostatic hindrance.²³ Fig. 5h depicts the pH effect on enzyme loading, and these results along with the surface charge data confirm the enhancing effect of a low pH immobilization environment.⁴

3.4. Optimization via the desirability function

In numerical optimization, the desired goal should be selected for each factor and response from the menu. The possible goals should be: maximize, minimize, target, within range, none (for responses only) and set to an exact value (factors only). Minimum and maximum levels must be imported for each factor included. Weights can be assigned to different goals to adjust the shape of its particular desirability function. The goals are combined into an overall desirability function. Desirability is an objective function that ranges from zero outside of the limits, to one at the goal which is maximized by the program.²⁴

Due to the curvature in the response surfaces and their combination in the desirability function, the existence of more than two maxima is possible. Starting from several points in the design space increases the possibility of finding the best local maximum.^12,25

In this method, first each response (y_i) is converted to an individual desirability function (d_i), and then the individual desirabilities are combined to give the overall desirability (D). The overall desirability function D is defined as the weighted geometric average of the individual desirability (d_f) according to eqn (2).^12,26

where n is the number of studied responses in the optimization process. A multiple response method was applied for the optimization of any combination of five goals, namely initial enzyme concentration, GA concentration, immobilization time, solution ionic strength and pH.

The numerical optimization seeks to maximize the desirability function.^27–29 All of the effective factors were set for maximum desirability. The goal of immobilized enzyme loading and the activity of the obtained biocatalyst could be chosen in different ways. The results of choosing different goals and values of importance for the amount and activity of the immobilized enzyme are reported in Table 2.

Table 2 Results of different numerical optimization conditions

Goal		Importance		Factor level					Response		Desirability
Imm. (mg gMNP^−1)	Act. (%)	Imm. (mg gMNP^−1)	Act. (%)	Cenz (ppm)	CGA (%)	Time (h)	pH	Cbuffer (mM)	Imm. (mg gMNP^−1)	Act. (%)
In range	Maximize	3	3	1000	0.50	3.9	5.7	100	554.84	23.99	0.930
Maximize	Maximize	3	3	1000	4.80	4	4.6	100	665.87	21.81	0.981
Maximize	Maximize	3	4	1000	1.61	3.8	4.6	99.89	595.12	23.84	0.979
Maximize	Maximize	3	5	994.43	0.59	4	4.6	99.99	556.41	24.83	0.979

---

As has been discussed in Section 3.3, the activity of the immobilized enzyme does not increase necessarily with enzyme loading. Therefore, the choices of the goal in the case of biocatalyst activity and enzyme loading, and also their importance values, have a significant effect on the optimum conditions. As the aim of this work is to obtain a highly active and stable biocatalyst, the goal of activity during optimization should be maximized. Also, from a practical point of view, biocatalysts with higher catalyst loadings are desired. Accordingly the immobilized enzyme loading should be maximized as well; but with lower importance in comparison to activity. According to Table 2, the changes in the importance of enzyme activity, as it was expected only leads to GA concentration change. The increase of GA concentration results in higher immobilization, hence higher activity; however in some cases, GA concentration increase leads to the crosslinking of immobilized enzymes and activity decrease via enzyme rigidity. Considering the aim of the immobilization process, to obtain a biocatalyst with high activity and enzyme loading, an immobilization condition with an activity importance value of 5 has been chosen as the optimum point. 3D surface plot desirability of the combined effects of the initial enzyme concentration and GA concentration for the optimal activity of the immobilized enzyme were shown in Fig. 6.

Fig. 6 Response surface plot of the interaction of enzyme and GA concentrations and their effect on desirability.

The best local maximum was obtained at an initial solution pH of 4.6, initial enzyme concentration of 994.43, GA concentration of 0.59, a time of 4 and an ionic strength of 99.99. At these optimal conditions the enzyme immobilization and activity were measured to be 556.41 mg g_MNPs⁻¹ and 24.83% respectively at a desirability value of 0.979. The value of the obtained desirability shows that the D function represents the experimental model under the desired conditions.

In order to verify the developed model, some of the experiments were carried out under optimum conditions and the obtained data was compared with the predicted results from the model. The obtained experimental data for the optimum sample (activity = 23.92%) showed acceptable fit with the model prediction (activity = 24.8%), confirming model validity.

3.5. Results of stability tests

3.5.1. Thermal and pH stability. Fig. 7a presents the relative activity of the free and immobilized enzyme in the temperature range of 30–80 °C. The maximum activities for both free and immobilized enzymes were observed at 30 °C. The relative activity loss due to temperature increase (above 60 °C) was approximately 80% for the immobilized enzyme and 100% for the free one. As can be seen in the figure, the immobilized amylase was more stable than the free amylase.

Fig. 7 Temperature (a) and pH (b) stabilities of the immobilized enzyme (solid line) and free (dashed line).

The increase in stability could be the result of an improvement in enzyme rigidity through covalent immobilization.^7,30

The effects of pH on the free and immobilized α-amylase were studied in the range of 4.0–8.0. As can be seen in Fig. 7b, the maximum activity of the free and immobilized enzyme occurs at pH 6. According to the results, pH variations have less effect on the immobilized enzyme activity in comparison to the free enzyme. The figure shows that the activity of the immobilized enzyme is higher than 80% in a wide range of pH compared to the free enzyme; the higher stability of the immobilized biocatalyst corresponds to the conformational rigidity and the diffusional limitations of the immobilized molecules.³¹

3.5.2. Storage stability and reusability. The residual activity as a function of the number of cycles is presented in Fig. 8a. According to the results, the optimum biocatalyst kept nearly 40% of its initial activity after six cycles. The immobilized enzyme was not released during the experiments, so the activity decline could be related to the conformational changes in the immobilized enzyme structure.³¹

Fig. 8 Catalyst recycling of the immobilized enzyme (a), and storage stability of the immobilized (solid line) and free (dashed line) α-amylase (b).

Fig. 8b illustrates the relative activity of immobilized and free α-amylase, stored in 20 mM phosphate–citrate buffer (pH = 7) at 4 °C. The activity was measured for 12 days. The amounts of remaining relative activity of the immobilized and free enzyme were about 50 and 23% respectively after 12 days. It confirms that α-amylase becomes more stable after immobilization.

3.6. Kinetic parameters

The Lineweaver–Burk plot was generated using the results of the free and immobilized enzyme (at optimal conditions) at pH = 6 and 25 °C (Fig. 1-S†).

The K_m values for the free and immobilized enzymes were calculated as 0.092 and 0.053 mM, respectively. The lower K_m values represent the higher affinity of the enzymes to substrates. According to the results, the affinity of α-amylase to its substrate was increased by immobilization. It could be attributed to the nonporous structure of MNPs that forces enzyme molecules to be expanded over the MNP surface through proper orientation, leading to highly available active sites⁵ The values of V_max for free and immobilized enzymes were found to be 7.96 and 5.36 μmol (mg min)⁻¹, respectively. Since enzymatic reactions have high reaction rates at enzyme active sites, the mass transfer of the substrate to the active site is the rate determining step in these reactions. Hence, the decline in V_max after immobilization corresponds to the mass transfer limitation of the diffuse layer around the biocatalyst particle.³¹

Table 3 compares the results of the current study with activity, enzyme loading, and biocatalytic kinetics of recent works regarding amylase immobilization on magnetic nanoparticles. A significant enzyme loading and activity enhancement are observed. This multiple fold enzyme loading increase could correspond to the optimization of the effective factors in the immobilization process; leading to the immobilization of the enzyme via different physical and covalent mechanisms. The multifunctional surface of the support, containing a high concentration of amino groups and activated by GA, increased the amylase loading capacity of the support considerably. Also, a remarkable decrease of K_m indicates the affinity increase of α-amylase to the substrate.

Table 3 Optimized properties of immobilized α-amylase in the present and other studies

Carrier	Loading (mg g^−1)	Activity (%)	Km		Vmax		Ref.
			Free	Immobilized	Free	Immobilized
Gum acacia stabilized Fe3O4 MNPs	0.6	—	2.2 mg ml^−1	2.9 mg ml^−1	3.4 U ml^−1	5.3 U ml^−1	1
Silica coated Fe3O4 MNPs	11.5	19.6	5.0 mg ml^−1	2.5 mg ml^−1	0.21 U mg^−1	0.78 U mg^−1	2
Dye attached magnetic Fe3O bead	401	10.5	0.780 mg ml^−1	0.785 mg ml^−1	4.09 U mg^−1	5.98 U mg^−1	4
Silica coated Fe3O4 MNPs	15	—	6.27 mM	4.77 mM	2.44 U mg^−1	11.58 U mg^−1	5
Modified Fe3O4 MNPs	556.41	24.8	0.092 mM	0.053 mM	7.96 U mg^−1	5.36 U mg^−1	This study

---

4. Conclusions

Aiming for an optimum immobilization route of α-amylase on MNPs, factors impacting the immobilization process and their interactions were investigated using a CCD. The initial enzyme concentration, solution pH, GA concentration, ionic strength and time significantly influenced the amount and activity of the immobilized enzyme. These were optimized for the maximum efficiency of the immobilized enzyme by applying a desirability function. A dimensionless individual desirability value of 0.979 indicates that the estimated function properly represents the experimental model under the desired conditions. The optimum immobilized α-amylase demonstrates remarkable activity, enhancement of pH and temperature stabilities, as well as increased reusability, storage time and substrate affinity.

References

1. V. Swarnalatha, R. A. Esther and R. Dhamodharan, J. Mol. Catal. B: Enzym., 2013, 96, 6–13.
2. A. Mukherjee, T. Kumar, S. Rai and J. Roy, Biotechnol. Bioprocess Eng., 2010, 15, 984–992.
3. M. J. Khan, Q. Husain and A. Azam, Biotechnol. Bioprocess Eng., 2012, 17, 8.
4. N. Tüzmen, T. Kalburcu and A. Denizli, J. Mol. Catal. B: Enzym., 2012, 78, 16–23.
5. N. Sohrabi, N. Rasouli and M. Torkzadeh, Chem. Eng. J., 2014, 240, 426–433.
6. K. Singh and A. M. Kayastha, J. Mol. Catal. B: Enzym., 2014, 104, 75–81.
7. F. Eslamipour and P. Hejazi, J. Mol. Catal. B: Enzym., 2015, 119, 1–11.
8. T. N. Nwagu, B. Okolo, H. Aoyagi and S. Yoshida, Process Biochem., 2013, 48, 1031–1038.
9. S. J. Kalil, F. Maugeri and M. I. Rodrigues, Process Biochem., 2000, 35, 539–550.
10. M. Amini and H. Younesi, Clean: Soil, Air, Water, 2009, 37, 10.
11. M. Talat, A. K. Singh and O. N. Srivastava, Bioprocess Biosyst. Eng., 2011, 34, 647–657.
12. M. A. Bezerra, R. E. Santelli, E. P. Oliveira, L. S. Villar and L. A. Escaleira, Talanta, 2008, 76, 965–977.
13. M. A. Islam, M. R. Alam and M. O. Hannan, Composites, Part B, 2012, 43, 861–868.
14. M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
15. G. L. Miller, Anal. Chem., 1959, 31, 426–428.
16. C. Guo, M. Yunhui, S. Pengfei and F. Baishan, Biochem. Eng. J., 2012, 67, 120–125.
17. K. Can, M. Ozmen and M. Ersoz, Colloids Surf., B, 2009, 71, 154–159.
18. H. Liu, J. Liu, B. Tan, F. Zhou, Y. Qin and R. Yang, Bioprocess Biosyst. Eng., 2012, 35, 1287–1295.
19. A. S. Campbell, C. Dong, F. Meng, J. Hardinger, G. Perhinschi, N. Wu and C. Z. Dinu, ACS Appl. Mater. Interfaces, 2014, 6, 5393–5403.
20. Y. J. Cho, O. J. Park and H. J. Shin, Enzyme Microb. Technol., 2006, 39, 108–113.
21. G. Bayramoğlu, S. Kiralp, M. Yilmaz, L. Toppare and M. Y. Arıca, Biochem. Eng. J., 2008, 38, 180–188.
22. O. Barbosa, C. Ortiz, A. Berenguer-Murcia, R. Torres, R. C. Rodrigues and R. Fernandez-Lafuente, RSC Adv., 2014, 4, 1583–1600.
23. S. Talekar, A. Joshi, G. Joshi, P. Kamat, R. Haripurkar and S. Kambale, RSC Adv., 2013, 3, 12485–12511.
24. H. Heidari and H. Razmi, Talanta, 2012, 99, 13–21.
25. M. Mourabet, A. El Rhilassi, H. El Boujaady, M. Bennani-Ziatni and A. Taitai, Arabian J. Chem., 2014 DOI:10.1016/j.arabjc.2013.12.028.
26. M. Fereidouni, A. Daneshi and H. Younesi, J. Hazard. Mater., 2009, 168, 1437–1448.
27. R. Singh, N. R. Bishnoi and A. Kirrolia, Bioresour. Technol., 2013, 138, 222–234.
28. M. Mourabet, A. El Rhilassi, H. El Boujaady, M. Bennani-Ziatni, R. El Hamri and A. Taitai, J. Saudi Chem. Soc., 2015, 19, 603–615.
29. R. Singh, N. R. Bishnoi and A. Kirrolia, Bioresour. Technol., 2013, 138, 222–234.
30. A. Pal and F. Khanum, Process Biochem., 2011, 46, 1315–1322.
31. G. Srivastava, S. Roy and A. M. Kayastha, Food Chem., 2015, 172, 844–851.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra26140f

---