Penicillin G acylase (PGA) is an important industrial biocatalyst that is widely used for the enzymatic production of 7- aminodeacetoxycephalosporanic acid and 6-amimopenicillanic acid, which are the major pharmaceutical intermediates for producing semisynthetic β-lactam antibiotics [1, 2]. However, the stability and reusability of PGA are major drawbacks that have restricted its wide industrial application [3, 4]. Therefore, as an industrial biocatalyst, it is extremely urgent to immobilize PGA on solid supports to obtain recoverable and stable heterogeneous biocatalysts. Many kinds of supports have been reported for the immobilization of PGA, including acrylic resins [5], aldehyde-agarose [6], gelatin-chitosan [7], magnetic polymer beads [8], functionalized silica [9] and mesoporous silica [10-12].

The optimal support material for the immobilization of PGA should meet the following requirements: (1) A high density of reactive groups on the support surface for multipoint grafting of enzymes [13] and (2) a large surface area to immobilize more enzymes on the support surface [14]. In general, the surface area can be increased by decreasing the particle size of the support. However, support materials with a small particle size are difficult to separate from the reaction mixture by conventional methods. For example, GAMM support with epoxy groups can meet the first requirement mentioned above and has been used as a support for covalent immobilization of PGA and glucose isomerase [15, 16]. However, with a decrease in the particle size, the reuse of GAMM support in the substrate by filtration is time-consuming and inconvenient. Magnetic separation is an attractive alternative to centrifugation or filtration. Therefore, among all supports, magnetic materials have attracted increasing attention for the immobilization of enzymes. Magnetic supports are mainly magnetic silica composites and polymer microspheres, which are prepared by encapsulating magnetic particles in a silica shell or in organic polymers [17, 18]. Yang et al. [19] prepared paramagnetic aldehyde-functionalized mesostructured cellular foams (PAMCFs) by grafting 3-aminopropyltriethoxysilane-modified Fe₃O₄ (NH₂-Fe₃O₄) nanoparticles on the outer surface of aldehyde-functionalized mesostructured cellular foams (AMCFs). PGA/PAMCFs-15 showed a high initial activity of 9563 U/g and retained 89.1% of its initial activity after it was recycled 10 times. Furthermore, PGA/PAMCFs can be easily recycled by a magnetic field instead of tedious separation by high-speed centrifugation that is usually used for mesoporous materials.

Herein, we prepared paramagnetic polymer microspheres as supports for the immobilization of PGA to combine the excellent properties of polymer microspheres with the unique paramagnetic properties of magnetic nanoparticles [20]. The magnetic nanoparticles were synthesized by a coprecipitation method, and then the paramagnetic polymer microspheres were synthesized by inverse suspension polymerization with glycidyl methacrylate (GMA), ally glycidyl ether (AGE) and acrylamide in the presence of silica-coated Fe₃O₄ nanoparticles. The structure, composition and morphology of the paramagnetic polymer microspheres were characterized by Fourier transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM), X-ray diffraction (XRD), vibrating sample magnetometry and nitrogen sorption. The initial activity, operational stability, thermal stability and pH stability of the immobilized PGA were investigated.

Penicillin G acylase (520 U/mL) was purchased from Zhejiang Haider Co. Ltd., Jinhua, China. Penicillin G potassium salt was purchased from CSPS Heibei Zhongrun Pharmaceutical Co. Ltd., Shijiazhuang, China. Tetraethyl orthosilicate (TEOS) and ammonium hydroxide (25 wt%) were bought from Shanghai Lingfeng Chemical Reagent Co. Ltd., Shanghai, China. Other reagents were of analytical grade and obtained from Sinopharm Chemical Reagent Co. Ltd., Shanghai, China.

Fe₃O₄ nanoparticles were prepared by a conventional coprecipitation method [21]: 0.0137 mol of FeCl₃·6H₂O and 0.0075 mol of FeSO₄·7H₂O were dissolved in 100 mL of deionized water in a three-neck flask under a nitrogen atmosphere and then the mixture was heated to 65 ℃ under mechanical stirring. Then, aqueous ammonia (25 wt%) was added dropwise into the solution until the pH value reached 11.0. After the mixture was stirred at 65 ℃ for 2 h, the Fe₃O₄ nanoparticles (designated as Fe₃O₄) were separated from the mixture by the aid of magnet, washed several times with deionized water until the pH decreased to 7.0 and then dried at 60 ℃ under vacuum for 12 h.

The SiO₂-coated magnetic Fe₃O₄ nanoparticles were prepared by the Stober method [22]. 1.5 g of Fe₃O₄ nanoparticles were added into the mixture composed of 120 mL of ethanol, 30 mL of deionized water and 3.75 mL of aqueous ammonia (25 wt%). The suspension was vibrated ultrasonically and then 2.1 g of TEOS was slowly added into the suspension under continuous stirring. After stirring at room temperature for 12 h, the SiO₂-coated magnetic Fe₃O₄ nanoparticles (designated as Fe₃O₄/SiO₂) were separated by the aid of a magnet, washed with deionized water until the pH reached 7.0 and then dried at 60 ℃ under vacuum for 12 h.

The paramagnetic polymer microspheres were prepared by the inverse suspension polymerization method [23]. Typically, 1.07 g of a mixture of Span-60 and Tween-20 (8:3, mass ratio) was added into 120 mL of the mixed solvent of n-heptane and tetrachloroethylene (3:1, volume ratio) in a glass flask. Then, 28 g of the mixture composed of Fe₃O₄/SiO₂ nanoparticles, GMA, AGE, N, N'-methylene-bis(acrylamide), methacrylamide and carboxamide (1:2:2:7:1:20, mass ratio) was added into the glass flask within 2 min with 0.55 g of azobisisobutyronitrile as an initiator under a nitrogen atmosphere. After the resulting mixture was reacted at 55 ℃ for 4 h, the produced paramagnetic polymer microspheres were separated with the aid of a magnet and washed with ethanol several times. Finally, the paramagnetic polymer microspheres were extracted with 50 mL of ethanol for 48 h and 50 mL of n-heptane for 48 h successively and dried at 60 ℃ under vacuum for 12 h.

0.1 g of paramagnetic polymer microspheres was added to 5 mL of PGA solution that was diluted with phosphate buffer (1 mol/L, pH = 7.5, V_buffer:V_PGA = 5.25:1). The mixture was incubated in a water bath at 28 ℃ for 12 h at a rotating speed of 150 r/min. Then, immobilized PGA was separated with the aid of a magnet and washed with phosphate buffer (0.1 mol/L, pH = 7.5) to remove the noncovalently coupled enzymes. The wet immobilized PGA was stored at 4 ℃ for the subsequent testing of the activity and operational stability.

Hydrolyzing penicillin G potassium salt (PGK) by immobilized PGA can produce the same molar amount of 6-amimopenicillanic acid and phenylacetic acid (PAA). Thus, the activity of immobilized PGA can be obtained by measuring the volume of NaOH solution consumed to titrate PAA in a given time. The procedure of testing the specific activity of immobilized PGA was as follows [15]: 2.5 g of PGK was dissolved in 47.5 mL of phosphate buffer (0.02 mol/L, pH = 8.0) in a water bath at 28 ℃, and then 0.15 g of wet immobilized PGA was added. The mixed solution was titrated with NaOH solution (0.1 mol/L) to maintain pH = 8.0. The hydrolysis procedure was continued for 10 min, and the volume of NaOH solution consumed during the last 5 min was recorded. The specific activity of wet immobilized PGA was calculated as follows:

V_NaOH is the volume of NaOH solution consumed (mL); C_NaOH is the concentration of NaOH solution (mol/L); m is the weight of wet immobilized PGA (g); t is the reaction time (min), i.e., 5 min.

The operational stability of the immobilized PGA was tested as follows: After testing the initial specific activity, immobilized PGA was separated with the aid of a magnet and washed with phosphate buffer (0.1 mol/L, pH = 7.5). Then, the activity of the immobilized recycled PGA was tested by the abovementioned method.

Powder XRD patterns were recorded on a Bruker AXS D8 Focus diffractometer operated at 40 kV, 40 mA (Cu K_α radiation, λ = 0.15406 nm). FT-IR spectra were recorded on a Nicolet Nexus 670 FT-IR spectrometer. Magnetic susceptibility was determined on a Lakeshore 7407 Vibrating Sample Magnetometer at room temperature. The morphology of the microspheres was observed by JEOL JSM-6360LV SEM. N₂ adsorption-desorption isotherms were obtained at –196 ℃ on a Micromeritics ASAP 2020M surface area and porosity analyzer. Prior to the measurements, the sample was vacuum-degassed at 100 ℃ for 12 h.

Table 1 shows the effect of the SiO₂ coating on the performance of the immobilized PGA on Fe₃O₄/GAMM supports. Immobilized PGA on Fe₃O₄/SiO₂/GAMM support showed a higher initial activity (430 U/g) and better operational stability (99%) than PGA on Fe₃O₄/GAMM support. The reason is that the magnetic Fe₃O₄ nanoparticles without a SiO₂ coating tend to aggregate and are unevenly dispersed into the polymer microspheres. As a result, magnetic Fe₃O₄ aggregates exist in the pores of the polymer microspheres, leading to a decrease in the number of sites for PGA immobilization and a decrease in the initial activity of immobilized PGA on the Fe₃O₄/GAMM support. Furthermore, some polymer microspheres had no magnetic Fe₃O₄ nanoparticles embedded in them, leading to a loss of immobilized PGA during the recycled experiments owing to the difficulty in separating immobilized PGA without magnetic properties from the reaction mixture. Consequently, the operational stability of the Fe₃O₄/GAMM support was lower than that of the Fe₃O₄/SiO₂/GAMM support. Moreover, owing to the high abundance of silanol groups on the surface of the SiO₂ coating, the SiO₂-coated magnetic Fe₃O₄ nanoparticles (Fe₃O₄/SiO₂) were easily and highly dispersed in the hydrophilic monomer phase during the preparation process of the paramagnetic polymer microspheres, which can overcome the disadvantage of the parent Fe₃O₄ nanoparticles [22].

Fig. 1 shows the performance of immobilized PGA on the paramagnetic polymer microspheres with different Fe₃O₄/SiO₂ contents. As shown in Fig. 1, the initial activity of the immobilized PGA decreased with increasing Fe₃O₄/SiO₂ content in the paramagnetic polymer microspheres. The reason is that more pores of the paramagnetic polymer microspheres were occupied by Fe₃O₄/SiO₂ nanoparticles with increasing the Fe₃O₄/SiO₂ content, leading to a decrease in the amount of PGA immobilized in the pores of the magnetic microspheres and a decrease in the initial activity of immobilized PGA on the Fe₃O₄/SiO₂/GAMM support. In contrast, the operational stability of immobilized PGA on the Fe₃O₄/SiO₂/GAMM support showed a reverse trend to the relative activity of immobilized PGA. In other words, the operational stability of immobilized PGA increased with increasing Fe₃O₄/SiO₂ content in the paramagnetic polymer microspheres. As shown in Fig. 1, when the mass content of Fe₃O₄/SiO₂ in the paramagnetic polymer microspheres was 2.5%, the relative activity of the immobilized PGA decreased to 60% after it was recycled 10 times. However, the relative activity of immobilized PGA hardly decreased after 10 cycles when the mass content of Fe₃O₄/SiO₂ in the paramagnetic polymer microspheres was increased to 7.5%. Therefore, from the viewpoint of the operational stability, the optimum mass content of Fe₃O₄/SiO₂ in the paramagnetic polymer microspheres was 7.5%. All subsequent characterizations were carried out for paramagnetic polymer microspheres with a Fe₃O₄/SiO₂ mass content of 7.5%.

Fig. 2 shows the surface morphology of the paramagnetic polymer microspheres. As shown in Fig. 2, the paramagnetic polymer microspheres exhibited a spherical shape with a rough surface, and their diameters were in the range of 100–150 μm.

Fig. 3 shows FT-IR spectra of the paramagnetic polymer microspheres. The absorption band at 596 cm^–1 was assigned to the stretching vibration of the Fe–O bond of the parent Fe₃O₄ nanoparticles [24], and the absorption bands at 907 and 845 cm^–1 were assigned to epoxy groups. The absorption band at 1210 cm^–1 was attributed to the C–O–C asymmetric stretching vibration of the ester group of GMA, whereas the strong absorption band at 3425 cm^–1 was attributed to the N–H stretching vibration of acylamides (MBAA, MAA). In addition, the absorption bands at 1685, 1526 and 1455 cm^–1 were assigned to the stretching vibrations of C=O, C–N, N–H in MBAA and MAA, respectively [15]. Therefore, it can be concluded that the copolymer has been formed and coated on the surface of the magnetic microspheres.

Fig. 4 shows XRD patterns of Fe₃O₄, Fe₃O₄/SiO₂ and the magnetic polymer microspheres. The XRD patterns of all the samples exhibited six characteristic diffraction peaks assigned to (220), (311), (400), (422), (511) and (440), which were in agreement with that of the standard Fe₃O₄ crystals with a spinel structure [25]. The XRD pattern of Fe₃O₄/SiO₂ sample hardly changed compared with the Fe₃O₄ sample, which indicated that the crystal structure of Fe₃O₄ was completely retained after Fe₃O₄ was coated with SiO₂. Moreover, in addition to the diffraction peaks related to Fe₃O₄ crystals, a wide diffraction peak at 20°–30°, assigned to amorphous polymer, appeared in the XRD pattern of the paramagnetic polymer microspheres, which further confirmed the existence of copolymer on the surface of the magnetic microspheres.

Fig. 5 shows the magnetization curve of the paramagnetic polymer microspheres. The magnetic polymer microspheres exhibited an apparent superparamagnetism without any hysteresis at room temperature. The superparamagnetic property was also deduced from the low residual magnetization and coercivity values [26]. The saturation magnetization (Ms) of the paramagnetic polymer microspheres was 5.0 emu/g. As shown in Fig. 6, immobilized PGA on the paramagnetic polymer microspheres could be rapidly separated within 1 s with the aid of a magnet, which made the recycling operation more convenient.

The thermal stability, pH stability and reusability of the immobilized PGA are very important criteria for its industrial applications. To determine the thermal stability of the PGA immobilized on the paramagnetic polymer microspheres, the relative activities of free and immobilized PGA were tested after incubation in phosphate buffer (0.02 mol/L, pH=8.0) at different temperatures (30–65 ℃) for 1 h. Similarly, to determine the pH stability of immobilized PGA, the relative activities of free and immobilized PGA were tested after incubation at 28 ℃ in phosphate buffer at different pH values (4.0–11.0) for 1 h.

Fig. 7 shows the thermal stability and pH stability of free and immobilized PGA on the paramagnetic polymer microspheres. As shown in Fig. 7(a), although both the relative activities of free and immobilized PGA decreased with increasing incubation temperature, the thermal stability of the immobilized PGA was much better than that of free PGA. After it was incubated at 55 ℃ for 1 h, immobilized PGA could retain 91% of its initial activity, whereas free PGA only retained 8.1% of its initial activity. The increase in the thermal stability may be owing to reinforcement of the weak intermolecular forces and the prevention of PGA autolysis [27]. As shown in Fig. 7(b), both the relative activities of free and immobilized PGA first increased and then decreased with increasing pH value of phosphate buffer. The optimum pH values of phosphate buffer were approximately 8.0 and 7.0 for free and immobilized PGA, respectively. Furthermore, immobilized PGA showed a wider pH operating window than free PGA. These results indicated that the immobilization of PGA on paramagnetic polymer microspheres by reacting the amino groups of the PGA molecules with the epoxy groups of paramagnetic supports provided high thermal stability and pH stability in extreme environments [28].

Fig. 8 shows the operational stability of immobilized PGA on the paramagnetic polymer microspheres. The initial activity of the immobilized PGA was 430 U/g (wet) and it could retain 99% of its initial activity after it was recycled 10 times, indicating very good reusability.

Paramagnetic polymer microspheres with active epoxy groups were prepared by the inverse suspension polymerization method. SiO₂ played an important role in the polymerization process and its coating on the surface of Fe₃O₄ nanoparticles influenced the performance of the immobilized PGA. Immobilized PGA on the paramagnetic polymer microspheres showed a high initial activity of 430 U/g (wet) and retained 99% of its initial activity after it was recycled 10 times. The material could be conveniently recycled from the reaction medium with the aid of a magnet. Furthermore, compared with free PGA, the immobilization of PGA through the reaction of the amino groups of the PGA molecules with the epoxy groups of the paramagnetic polymer microspheres resulted in a high thermal stability and pH stability in extreme environments. Therefore, PGA immobilized on paramagnetic polymer microspheres is an excellent potential catalyst for industrial applications.