Potassium permanganate, barium chloride, hydrochloric acid, chloroacetic acid, ethanol, ammonium hydroxide aqueous solution (25%, by mass), iron(II) sulfate heptahydrate, hydrogen peroxide, sodium dihydrogen phosphate, disodium phosphate, glutaraldehyde, sulfuric acid, phenol, 4-aminoantipyrine (4-AAP) and sodium hydroxide were obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai, PR China). Iron(III) chloride hexahydrate, polyethylene glycol (PEG), graphite and sodium nitrate were bought from Aladdin Industrial Corporation (Shanghai, PR China). The 6arm-PEG-NH₂ was supplied by Shanghai Jinpan Biotech Co., Ltd. (Shanghai, PR China). Horseradish peroxidase (HRP) was bought from Shanghai Jinsui Biotechnology Co., Ltd (Shanghai, PR China). The 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDAC) was bought from Sangon Biotech Co., Ltd (Shanghai, PR China). All other reagents were of analytical grade and used without any further treatment. Deionized water was used in all the experiments.

Graphene oxide (GO) was synthesized using Hummers method (17). Masses of 0.5 g graphite and 0.5 g NaNO₃ were added to the concentrated H₂SO₄ (12 mol/L, 23 mL) solution and stirred at 600 rpm on a mechanical agitator in ice bath for 15 min. After slowly adding 4.0 g of KMnO₄, the mixture was transferred to a water bath heated to 40 °C and stirred (JJ-1; Xinrui Apparatus Factory, Jintan, PR China) at 600 rpm for 90 min. A volume of 50 mL of deionized water was added to the reaction system and stirred for 10 min. After adding 6 mL of hydrogen peroxide, the colour of the mixture changed from purple to golden. After standing for 24 h, the supernatant was decanted, the precipitate was collected by suction filtration and washed repeatedly with 5% hydrochloric acid until no precipitate was detected in the supernatant with BaCl₂. The precipitate was then washed with deionized water and centrifuged (TGL-16aR; Shanghai Anting Scientific Instrument Factory, Shanghai, PR China) at 8000×g until the pH of the supernatant reached pH=7. It was then placed in a cuvette and dried in a vacuum oven (DZF-6021; Shanghai Yiheng Instruments Co., Ltd, Shanghai, PR China) to constant mass. Next, to 1.0 g of the above product 0.4 g NaOH and 200 mL water were added and the mixture was ultrasonicated (KQ-250B; Kunshan Ultrasonic Instruments Co., Ltd, Kunshan, PR China) for 2 h. Then, 0.5 to 0.6 g of chloroacetic acid was added and ultrasonicated for 2 h. The solution was centrifuged at 8000×g, washed repeatedly with deionized water to neutral pH and dried in a vacuum oven to constant mass to obtain GO-COOH, which is rich in carboxyl groups on the surface.

Fe₃O₄ was synthesized on the surface of GO-COOH by coprecipitation (18, 19). A mass of 0.2 g of GO-COOH was dispersed in 100–150 mL of water, then 0.466 g of FeCl₃·6H₂O and 0.48 g of FeSO₄·7H₂O were added and ultrasonicated (KQ- -250B; Kunshan Ultrasonic Instruments Co., Ltd.) for 10 min. The temperature of the mixture was raised to 60 °C and 20 mL of aqueous ammonia (3.5 moL/L) were added and reacted for 60 min. The product was separated with a magnet, washed with deionized water to neutral pH and then dried to constant mass in an oven (DHG-9077; Taicang Jinghong Experimental Equipment Co., Ltd, Taicang, PR China) to obtain magnetic graphene oxide GO@Fe₃O₄.

According to a previous report (20), the improved method of 6arm-PEG-NH₂ grafting was applied in this study. A mass of 200 mg of GO@Fe₃O₄ together with 20 mL of 6arm-PEG- -NH₂ were added to 200 mL of deionized water and sonicated (KQ-250B; Kunshan Ultrasonic apparatus Co., Ltd.) for 1 h. After adding 40 mg of EDAC to the solution, the mixture was stirred for 2 h at room temperature and then 52 mg of EDAC were added. Stirring continued for 12 h, then the product was separated by an external magnetic field, washed with deionized water to neutral pH, and then dried under vacuum (DHG-9077; Taicang Jinghong Experimental Equipment Co., Ltd) at 40 °C.

The chemical structure of the magnetic material was determined by a Fourier transform infrared (FTIR) spectrometer (Nexus 470; Thermo Fisher Scientific, Madison, WI, USA) using KBr pellets, and the diffuse reflectance spectra were scanned over the range of 400–4000 cm^–1. The morphology of each step product was detected by transmission electron microscopy (Tecnai™ 12; FEI Company, Hillsboro, OR, USA). During the test, the sample was dispersed in ethanol and spotted onto a carbon-coated copper mesh and then dried at room temperature. The magnetization curves of GO@Fe₃O₄ and GO@Fe₃O₄@6arm-PEG-NH₂ were measured by vibrating sample magnetometer (VSM 7410; Lake Shore Cryotronics, Inc., Westerville, OH, USA) at room temperature. The X-ray photoelectron spectroscopy (XPS) measurements were performed using a hemispherical analyzer (ESCALAB™ 250Xi; Thermo Fisher Scientific, Waltham, MA, USA) equipped with Al Kα radiation as X-ray source (k=100 eV).

A mass of 100 mg of magnetic material was added to 100 mL of glutaraldehyde solution and shaken on vortex mixer (HB20; Hangzhou Longyang Scientific Apparatus Co., Ltd., Hangzou, PR China) at 50 °C and 200 rpm for 6 h. The magnetic material was separated from the solution with a magnet and dried in vacuum oven (DZF-6020; Shanghai Yiheng Scientific Apparatus Co., Ltd., Shanghai, PR China). The magnetic material was prepared as a suspension with 0.2 M phosphate-buffered saline (PBS; pH=7.0). The HRP solution was added to the suspension and mixed uniformly. The mixture was incubated at constant temperature for 120 min and the immobilized HRP was separated with a magnet. The concentration of the enzyme in the supernatant after separation was examined. The support material was rinsed twice with a buffer solution to remove free HRP. The concentration of residual HRP in the supernatant solution was detected using UV-Vis spectrophotometer (UV-2450; Shimadzu, Tokyo, Japan), according to the Bradford protein assay method (21). The loading mass of HRP (Q/(mg/g)) can be calculated by the following equation:where m₁ is the protein mass (mg) added to the solution, m₂ is the protein mass (mg) in the remaining supernatant, and m is the mass (g) of the support material.

Enzyme activity was assayed using the traditional Worthington method (22). A volume of 1.4 mL of 4-AAP, 100 μL of immobilized HRP solution and 1.5 mL of H₂O₂ solution were added to the cuvette at 25 °C. The absorbance was measured at 510 nm with a UV-Vis spectrophotometer (UV-2450; Shimadzu) each minute until the increase in the absorbance per minute remained the same. The enzyme activity (EA/(U/mL)) was calculated according to the following equation:where A_{510 nm} denotes the increase of absorbance per minute at 510 nm, V_t is the total volume of the reaction solution (3 mL in this experiment), V_e is the volume of free or immobilized enzyme HRP (0.1 mL in this experiment), 10 is the number of dilutions of the sample and 0.685 is the increase of the absorbance of a unit of horseradish peroxidase in 1 minute.

The reaction solution (pH=7.0) contained 100 mg/L phenols, 0.2 mg/L immobilized enzyme, 200 mmol/L PBS, H₂O₂ (the molar ratio of H₂O₂/phenol was set at 1) and PEG (the mass ratio of PEG/phenol was set at 0.4) (23). After completion of the reaction, it was quenched by the addition of concentrated sulfuric acid. The solution was made up to 10 mL with PBS (pH=7.0). The immobilized HRP was collected by magnet adsorption. The supernatant was filtered and the phenol concentration was measured. The equation for calculating the degradation rate of phenols (DR_p/%) is as follows:where γ¹_p is the concentration of phenols in the initial solution (mg/L), γ²_p is the concentration of phenols in the remaining liquid (mg/L), and V is the volume of the reaction solution (L).

Fig. 1 shows the preparation process of the magnetic material. First, the graphite was oxidized to monolayer graphene oxide. Then the surface of the graphene oxide was carboxylated and the carboxylation of GO with Fe₃O₄ was carried out on this basis. Then, the 6arm-PEG-NH₂ was grafted onto the surface of the carboxylated graphene oxide by the reaction of the carboxyl group with the amino group. After that, the amino group of the support material and the amino group of the HRP were connected by glutaraldehyde to form an immobilized HRP.

Fig. 2a shows the FTIR spectra of GO, GO@Fe₃O₄ and GO@-Fe₃O₄@6arm-PEG-NH₂. GO gave several strong characteristic peaks corresponding to the oxygen functionalities, such as the CO stretching vibration peak at 1732 cm^–1, O-H bending and stretching vibration in the C-OH groups at 1404 and 3415 cm^–1. The absorption peak at 573 cm^–1 for GO@Fe₃O₄ (Fig. 2a) was attributed to the iron oxide on the GO surface. The absorption peaks at 1083, 1579 and 2879 cm^–1 (Fig. 2a) were assigned to the vibrations of C-O, N-H and C-H in 6arm-PEG-NH₂ respectively, which demonstrated that 6arm-NH₂ had been grafted onto GO@Fe₃O₄. Compared to the spectrum for GO@Fe₃O₄@6arm- -PEG-NH₂, the spectrum for GO@Fe₃O₄@6arm-PEG-NH₂/HRP had additional peaks at 1400−1600 cm⁻¹ for −CONH, and 2800−3000 cm⁻¹ for −CH₂ and −CH₃. Moreover, the GO-Fe₃O₄-6arm-PEG-NH₂/HRP spectrum shows peak shift (1400−1600 cm⁻¹) compared to HRP spectrum, indicating that HRP was immobilized via covalent bonding. Fig. 2b shows the X-ray diffraction (XRD) patterns of GO, GO@Fe₃O₄ and GO@Fe₃O₄@6arm- -PEG-NH₂. The diffraction peak characteristic of GO appeared at around 2θ=10°. Six peaks at 30.1, 35.5, 43.1, 53.4, 57.0 and 62.6° correspond to crystal values of Fe₃O₄ given by the International Centre for Diffraction Data (ICDD) (24) file (card no. 19-629) indexed as (220), (311), (400), (422), (511), (440), respectively. The Fe₃O₄ crystal form did not change during the reaction.

The specific magnetization values (σ) of the samples GO@-Fe₃O₄ and GO@Fe₃O₄@6arm-PEG-NH₂ were 39.61 and 30.80 A·m²/kg, respectively (Fig. 2c). There was no significant residual magnetic and coercive force at room temperature, indicating that it had superparamagnetic properties. The specific magnetization values were high enough to allow immobilized HRP to be quickly and efficiently separated from the solution.

Components of the composite were measured by thermogravimetric analysis (TGA), and the results are in Fig. 2d. The organic components were decomposed and inorganic components were retained with the increase of temperature during TGA. The 74.5% (by mass) loss of GO microspheres was attributed to the mass ratio of free water and oxygen-containing group of the GO. After binding with Fe₃O₄, the 23.1% loss of GO@Fe₃O₄ was assigned to the physically adsorbed water and oxygen-containing group of the GO. When the 6arm-PEG-NH₂ was introduced, the mass loss of the composite microspheres was about 52.7%, which was much higher than of GO@Fe₃O₄. The first mass loss (14.4%) at temperatures up to 200 °C was due to the evaporation of the physically adsorbed water or solvent, and the second major mass loss (38.3%) at temperatures from 200 to 800 °C was due to the decomposition of the polymer component 6arm-PEG- -NH₂ and GO of the multi-armed magnetic graphene oxide composite.

Fig. 2e shows the X-ray photoelectron spectroscopy (XPS) spectrum of the GO@Fe₃O₄@6arm-PEG-NH₂ magnetic material. The spectrum of C1s was dominated by a peak around 286.4 eV, which is assigned to graphene carbon, and the spectrum of Fe2p can also be found in the XPS spectrum, indicating the successful synthesis of GO@Fe₃O₄. The broad peaks around 711.2 and 734.3 eV were typical for magnetite, confirming the formation of Fe₃O₄. The spectrum of N1s was dominated by a peak around 396.6 eV, which is associated with 6arm-PEG-NH₂.

To explore the secondary structural changes of HRP after immobilization on GO@Fe₃O₄@6arm-PEG-NH₂, the circular dichroism (CD) spectra of HRP and GO@Fe₃O₄@6arm-PEG- -NH₂/HRP were recorded from 190 nm to 250 nm (Fig. 2f). Two negative peaks at 208 and 222 nm are observable, which are typical for α-helical structure of the enzyme, a common characteristic of all HRP, while a positive peak at around 196 nm was assigned to β-sheet structure. Compared with HRP, the secondary structure of GO@Fe₃O₄@6arm-PEG-NH₂/HRP did not significantly change. This shows that the immobilization had no drastic effect on the secondary structure of HRP.

Fig. 3 shows transmission electron microscopy (TEM) images of the acquired magnetite support material. TEM image of GO-COOH (Fig. 3a) shows that it has lamellar structure with obvious folds on the surface, which could provide attachment sites for Fe₃O₄ nanoparticles and prevent the mutual agglomeration of nanoparticles, as reported in the literature (25). Fe₃O₄ nanoparticles can be uniformly dispersed on the surface of GO (Fig. 3b). Statistical analysis showed that the average size of Fe₃O₄ particles was (20±5) nm. Comparingly, TEM image of GO@Fe₃O₄@6arm-PEG-NH₂ (Fig. 3c) shows a large shadow, representing the surface of magnetic GO covered with a layer of polymer, which indicated the successful modification of 6arm-PEG-NH₂.

The amino group of the enzyme molecule and the amino group of the magnetic material GO@Fe₃O₄@6arm-PEG-NH₂ can be ligated by glutaraldehyde, but the high concentration of glutaraldehyde can also lead to inactivation of the enzyme or cause self-crosslinking of the enzyme molecule. Therefore, it is necessary to find the appropriate concentration of glutaraldehyde. Fig. 4a shows that when the glutaraldehyde was not added, the loading of HRP was 13.15 mg/g, which indicated that the physical adsorption of the enzyme to the support material was weak. In order to immobilize the enzyme more firmly, the binding force between the two can be enhanced by the addition of a crosslinking agent. When the concentration of glutaraldehyde was increased from 0.05 to 0.10 mol/L, the loading of HRP increased from 95.17 to 143.13 mg/g. However, when glutaraldehyde concentration continued to rise, the loading of HRP was almost unchanged. With the increase of glutaraldehyde concentration, the phenomenon of self-crosslinking between the enzyme molecules was exacerbated and the enzyme molecules did not bind to GO@-Fe₃O₄@6arm-PEG-NH₂ anymore, resulting in a slight decrease in the loading of enzyme. Therefore, the optimal concentration of glutaraldehyde was 0.10 mol/L.

Fig. 4b shows the effect of the concentration of support material on the immobilization; the loading of the enzyme first increased, and then decreased. The average loading of HRP was 187.47 mg/g when the concentration of support material was lower than 7 mg/L, which may be due to the fact that HRP was almost completely immobilized at these concentrations. Therefore, the optimal concentration of support material was determined to be 7 mg/L.

Different ambient temperatures, ranging from 20 to 50 °C, were tested. Fig. 4c shows that with the increase of ambient temperature, the loading of HRP increased. When the temperature was 35 °C, the loading of HRP reached the maximum value of 189.99 mg/g. However, when the temperature was higher than 35 °C, the loading of HRP decreased slightly. Therefore, 35 °C was chosen as the optimal ambient temperature.

Finally, different reaction times were investigated. Fig. 4d shows that the loading of immobilized HRP increased with time ranging from 30 to 120 min. It reached a steady value of 180.47 mg/g after 120 min. It can be seen that the immobilization reaction was almost balanced at 120 min. Thus, the optimal reaction time was set at 120 min.

To sum up, under the optimal immobilization conditions (concentration of glutaraldehyde 0.1 mol/L, concentration of support material 7 mg/L, temperature 35 °C and incubation time 120 min), the average loading of HRP was 186.34 mg/g; relative standard deviation (RSD)=0.879%. However, the activity of the immobilized HRP was about 45.59 U/mg, which equaled only 76.9% of initial activity. In addition, the loading of immobilized HRP was comparably higher than previously reported supports for HRP (Table 1 (26–30)). All the above shows that the loading of immobilized HRP on GO@Fe₃O₄ can be effectively improved by grafting 6arm-PEG.

Fig. 5 shows the effect of various parameters on enzyme activity. The optimum pH of immobilized and free HRP was determined using different buffers with pH values of 5.5–8.5 and the results are shown in Fig. 5a. It can be seen that catalytic activities of immobilized HRP and free HRP reached the highest values at pH=7.5. This is consistent with the maximum activity of free HRP reported by Zhai et al. (30). In addition, when the pH was higher or lower than 7.5, the relative activity of the enzyme decreased. At higher pH, the relative activity of free HRP decreased more than of the immobilized HRP, which showed that the tolerance of immobilized HRP to the increased pH of the environment slightly improved. The pH stability of HRP can be improved by loading the HRP onto the surface of GO@Fe₃O₄@6arm-PEG-NH₂ (31).

The optimum catalytic temperature of free and immobilized HRP was found to be 25 to 70 °C (Fig. 5b). The relative activity of the immobilized and free HRP increased with the increase of temperature and reached the maximum at 55 °C. The change of activity may be due to the gradual increase of the kinetic energy of the molecule with the increase of reaction temperature, and the increase of the collisional frequency between HRP and the substrate. As the temperature continued to rise, the relative activity of the enzyme began to decline, especially of the free HRP. This is attributed to too high temperature, causing the polypeptide that makes up the enzyme to deform, which reduces its activity.

The kinetic parameters (K_m and v_max) of free HRP changed after immobilization on the support material. Specifically, the apparent K_m values of the free and immobilized HRP were 1.4525 and 0.6016 mmol/L, respectively. After immobilization, the v_max decreased slightly from 1230 to 1183 μM/s. The lower K_m values of immobilized HRP suggested a higher affinity towards the phenolic substrate after immobilization. The change in the affinity of HRP to its substrate was maybe caused by structural changes in the enzyme.

The tolerance of immobilized HRP to temperature was investigated. Fig. 5c shows that the temperature increase from 30 to 80 °C caused gradual decrease of the enzyme activity. This may be due to the deformation of the polypeptide, which leads to inactivation of the HRP at high temperature. The thermal stability of the immobilized HRP greatly improved. Its activity was much higher than that of the free HRP at high temperature (80 °C).

Stability of free and immobilized HRP during cold storage (4 °C) is shown in Fig. 5d. Under the same storage conditions, the activity of immobilized HRP decreased more slowly than that of the free HRP. The relative activity of immobilized HRP after 30 days was 85.5%, which was significantly higher than of free HRP (42.3%). After 60 days, free HRP was almost completely inactivated (10.2%), while the activity of immobilized HRP was still at a high level (72.5%). It was concluded that the stability of HRP was enhanced by immobilization, which was in line with the thermal stability data (Fig. 5c). It was also concluded that the enzyme immobilized on GO@Fe₃O₄@6arm- -PEG-NH₂ was more stable and had a longer shelf-life than other types of support materials (Table 2 (32–34)).

Fig. 5e shows the reusability of immobilized HRP in eight catalysis runs. The activity of the immobilized HRP decreased with the increase of recycling number, and it was over 68.1% of its initial value after eight cycles. The relative activity of immobilized HRP reached 76.0% after five cycles, and was significantly higher than 41% on GO@Fe₃O₄ as supporting material reported previously (35). This activity loss might be attributed to the accumulation of the products during the enzymatic reaction on the surface of the immobilized HRP, which blocked some of its reaction sites.

Fig. 6 shows the effect of different reaction parameters, including concentrations of phenols, molar ratio of H₂O₂ and phenols, concentration of immobilized HRP and reaction temperature on the degradation of phenols. Phenol removal efficiency increased rapidly with the increase of phenol concentration up to 100 mg/L, when it it reached the maximum value of 41.26% (Fig. 6a). However, when the concentration of phenols increased further, removal efficiency began to decline. This was due to the fact that the catalytic reaction rate was accelerated at first with the increase of substrate concentration, and when the substrate concentration was increased further, the aggregation of the product inhibited substrate degradation. Therefore, the phenol concentration of 100 mg/L was used as the optimum in the subsequent removal reaction.

The removal efficiency of phenols increased linearly with the increase of molar ratio of H₂O₂ to phenols up to 2.0, when it reached 63.13%, and then decreased (Fig. 6b). The reason for this could be that too much H₂O₂ causes excessive oxidation of the iron ions at the centre of HRP activity and affect the transmission of electrons, which may prevent catalytic reaction. Therefore, the optimal molar ratio of H₂O₂ to phenol was determined to be 2.0.

Fig. 6c shows that phenol removal efficiency increased rapidly with the increase of the concentration of immobilized HRP up to 0.2 mg/L, when it started to decrease. Based on the above findings, the optimal concentration of immobilized HRP was found to be 0.2 mg/L.

The effect of different reaction temperatures on the phenol removal efficiency was studied (Fig. 6d), and we concluded that it was not significant. Phenol removal efficiency decreased slightly with the increase of reaction temperature. Considering the thermal stability of immobilized HRP, optimal reaction temperature was determined to be 20 °C.

The degradation of phenols (100 mg/L) by the catalytic activity of immobilized HRP, free HRP, GO@Fe₃O₄ and GO@-Fe₃O₄@6arm-PEG-NH₂ in the presence of H₂O₂ (1.59 mmol/L) at pH=7.5 and 20 °C was investigated. Fig. 6e shows that the degradation rate of phenols by immobilized HRP was 85.7% after 10 min, while by the free HRP, GO@Fe₃O₄ and GO@-Fe₃O₄@6arm-PEG-NH₂, it was 45.6, 13.5 and 14.1% respectively. The relative activity of the pure carrier material was very weak because there was only non-specific adsorption effect between the carrier and the phenols. It indicated that the degradation of phenols by immobilized HRP was significantly better than by free enzyme and support material alone, due to the synergistic effect between the HRP and the support material, which is consistent with the previous literature report (26).