Synthesis, characterization and application of core–shell magnetic molecularly imprinted polymers for selective recognition of clozapine from human serum

Abstract

In this article, a magnetic molecularly imprinted polymer (MMIPs) based on Fe₃O₄@SiO₂ has been synthesized for simply extraction of clozapine (CLZ) from human serum. The MIPs were coated on the Fe₃O₄@SiO₂–NH₂ surface by the copolymerization of methacrylic acid with ethylene glycol dimethacrylate; and clozapine as a template molecule. The properties of the obtained Fe₃O₄@SiO₂–MIPs were characterized by Fourier transform infrared spectroscopy, X-ray diffraction, transmission electron microscopy, scanning electron microscopy, dynamic light scattering, energy dispersive X-ray analysis (EDX), a vibration sample magnetometer and rebinding experiments. The Fe₃O₄@SiO₂–MIPs showed a highly improved imprinting capacity, a fast adsorption equilibrium, and significant selectivity in molecularly imprinted solid-phase extraction of CLZ from human serum. In addition, MMIPs were regenerated and their adsorption capacity in the eighth use showed about an 6.67% loss in clozapine solution. Also, the intra and inter-day precision values were less than 6 and 4%, respectively. These results suggest Fe₃O₄@SiO₂–MIPs may be used for selective extraction and analysis of CLZ in human serum.

---

1. Introduction

Molecular imprinting is a powerful technique for creating specific binding sites in molecularly imprinted polymers (MIPs).^1–4 MIPs are obtained by the polymerization of functional monomers using a cross-linker around a template. The removal of the template leaves behind recognition sites with function and shape complementary to the template.^5–7 Compared to enzymes, antibodies, or biological molecules, MIPs show various inherent advantages such as (1) MIPs provided good physical and chemical stability; (2) MIPs can be used in harsh chemical environments without loss of binding properties; (3) molecularly imprinted polymers (MIPs) are inexpensive to produce, reusable, and are applicable to a number of different operating conditions. Furthermore, the preparation of MIPs is quite simple and economical.^7,8 These materials have been used for a wide variety of applications, including in the fields of catalysis,⁹ chemical sensors,¹⁰ food analysis,¹¹ solid-phase extraction (SPE),^12–15 chromatography,^16,17 environmental analysis,¹⁸ pharmaceutical analysis¹⁹ and drug controlled release.²⁰ MIPs have been prepared by bulk polymerization,^21–24 precipitation polymerization,^25–27 membrane polymerization,^28,29 multi-step swelling and polymerization^30–32 or surface-grafting polymerization.^33–35 MIPs as SPE sorbents have attracted considerable attention for being able to selectively recognize the target molecules from a mixture of chemical species. With the development of molecular imprinted-SPE (MISPE), magnetic separation technology, in which polymers are prepared by fabricating the MIP on the surface of a magnetic substrate, has received considerable attention in recent years for its potential application in extraction and separation.^36–39 Due to the high magnetic susceptibility of magnetic nanoparticles (MNPs), magnetic molecularly imprinted polymers (MMIPs) can be easily collected and separated by an external magnetic field without additional centrifugation or filtration, which makes separation easier and faster.⁴⁰

Recently, magnetic core–shell materials, which combine features of cores and various functional shells for different applications, have gained much attention and undergone intensive investigation for their unique magnetic properties, good stability,⁴¹ biological safety,⁴² low cytotoxicity,^43,44 chemically modifiable surface,^45,46 potential applications in medicinal,^47,48 environmental,⁴⁹ optical,⁵⁰ chemical areas^51,52 and magnetic molecularly imprinted polymer.^53,54 For this purpose, Fe₃O₄@SiO₂ nanoparticles is an excellent candidate as the supporting material because of its reactivity with various coupling agents, compatibility, reliable chemical stability, and the inherent properties of the magnetic element, which makes the resulting MIP magnetically susceptible and therefore, be easily separated by an external magnetic field after adsorption and recognition.^54,55

Clozapine is an effective antipsychotic drug treating schizophrenia with high toxicity.⁵⁶ Unlike traditional antipsychotics, clozapine has a low propensity for extrapyramidal side effects and minimal effect on prolactin secretion. Despite its therapeutic potential, wider use of clozapine has been limited by the high risk of agranulocytosis occurring in 1–2% of the patients.⁵⁷ Several analytical methods have been described for the quantification of clozapine and its metabolites in human plasma or serum, based on a variety of detection and separation techniques. Owing to its importance, researchers used many analytical methods to study it, including capillary zone electrophoresis,⁵⁸ high-performance liquid chromatography,⁵⁹ spectrophotometry,⁶⁰ mass spectrometry, etc.⁶¹

In this work, we prepared magnetic Fe₃O₄@SiO₂–MIPs with uniform core–shell structure by combining surface imprinting and nanotechniques. First, Fe₃O₄@SiO₂ nanosphere core–shell was synthesized using nano Fe₃O₄ as the core, TEOS as the silica source and PVA as the surfactant. Then, Fe₃O₄@SiO₂–NH₂ was prepared by coupling the aminosilica groups on the Fe₃O₄@SiO₂ surface through (3-aminopropyl) triethoxysilane (APTES). The MIPs were coated on the Fe₃O₄@SiO₂–NH₂ surface by the copolymerization of the amino end groups with functional monomer, methacrylic acid (MAA), using a cross-linking agent, ethylene glycol dimethacrylate (EGDMA); an initiator, azobisisobutyronitrile (AIBN); and a template molecule, clozapine.

The characterization, binding capacity, kinetics, selectivity and regeneration of these Fe₃O₄@SiO₂–MIPs particles were investigated. Finally, Fe₃O₄@SiO₂–MIPs was used as sorbent for extraction and determination of clozapine in human serum.

2. Experimental

2.1. Materials and physical measurements

Chemical reagents used in our experiments were purchased from the Merck Chemical Company and were used as received without further purification. Fourier transform infrared (FT-IR) spectra were obtained using a Shimadzu FT-IR 8300 spectrophotometer. X-ray powder diffraction (XRD) analysis was conducted on a Bruker AXS D8-advance X-ray diffractometer using Cu Kα radiation (λ = 1.5418). Scanning electron microscopy (SEM) images were obtained on Philips XL-30ESEM and transmission electron microscopy (TEM) images were obtained on a Philips EM208 transmission electron microscope with an accelerating voltage of 100 kV. The elements in the samples were probed by energy-dispersive X-ray (EDX) spectroscopy accessory to the Philips scanning electron microscopy (SEM). Magnetic characterization was carried out on a vibrating sample magnetometer (Meghnatis Daghigh Kavir Co., Iran) at room temperature and dynamic light scattering (DLS) were recorded on a HORIBA-LB550. Clozapine, diazepam, fluoxetine and phenytoin were purchased from Tehran Chemie (Tehran, Iran). The structures of chemicals used or assayed in this study are presented in Fig. 1.

Fig. 1 Structures of the chemicals used or assayed in this study.

2.2. Preparation of standard and calibration solutions

10 mg of CLZ was dissolved in 10 mL methanol to prepare stock solution. The standard solutions (6.1, 15.3, 30.6, 61.2, 153, 306 μM) used for rebinding tests were prepared from stock solution by dilution with ACN (acetonitrile). The standard solutions (2, 5, 10, 20, 50, 100 μg mL⁻¹) used for spiking calibration samples were prepared from the stock solution by dilution with water. 100 μL of standard solutions was added to 900 μL serum to obtain calibration standards (0.2, 0.5, 1, 2, 5, 10 μg mL⁻¹). The serum samples were frozen and stored in −20 °C.

2.3. Preparation of Fe₃O₄@SiO₂–MIPs

2.3.1. Synthesis of Fe₃O₄ NPs. The mixture of FeCl₃·6H₂O (1.3 g, 4.8 mmol) in water (15 mL) was added to the solution of polyvinyl alcohol (PVA 15 000) (1 g), as a surfactant, and FeCl₂·4H₂O (0.9 g, 4.5 mmol) in water (15 mL), which was prepared by completely dissolving PVA in water followed by addition of FeCl₂·4H₂O. The resultant solution was left to be stirred for 0.5 h in 80 °C. Then hexamethylenetetramine (HMTA) (1.0 mol L⁻¹) was added drop by drop with vigorous stirring to produce a black solid product when reaction media reaches pH 10. The resultant mixture was heated on water bath for 2 h at 60 °C and the black magnetite solid product was filtered and washed with ethanol three times and was then dried at 80 °C for 10 h.^62–64

2.3.2. Synthesis of amino-modified Fe₃O₄@SiO₂ NPs (Fe₃O₄@SiO₂–NH₂). The core–shell Fe₃O₄@SiO₂ nanospheres were prepared by a modified Stober method.^65–67 Briefly, Fe₃O₄ (0.50 g, 2.1 mmol) was dispersed in the mixture of ethanol (50 mL), deionized water (5 mL) and tetraethoxysilane (TEOS) (0.20 mL), followed by the addition of 5.0 mL of NaOH (10 wt%). This solution was stirred mechanically for 30 min at room temperature. Then the product, Fe₃O₄@SiO₂, was separated by an external magnet, and was washed with deionized water and ethanol three times and dried at 80 °C for 10 h.

Then, Fe₃O₄@SiO₂ (0.5 g) was added to the solution of 3-(triethoxysilyl)-propylamine (1 mmol, 0.176 g) in ethanol (5 mL) and the resultant mixture was under reflux for 12 h under nitrogen atmosphere. After refluxing, the mixture was cooled to room temperature, filtered by an external magnet and the product was washed with ethanol and water to remove unreacted species and dried at 80 °C for 6 h.^68,69

2.3.3. Preparation of the core–shell Fe₃O₄@SiO₂–MIPs and Fe₃O₄@SiO₂–NIPs. 1 g Fe₃O₄@SiO₂–NH₂, CLZ as the template and MAA as the functional monomer, were dissolved in 10 mL acetonitrile under ultrasonic vibration for 10 min, and then the mixture was placed in dark for 12 h to form a template–monomer complex.

Then, EGDMA as the cross-linker, and AIBN (0.2 g) as the initiator, were added. The reaction mixture was purged with nitrogen and stirred at 60 °C for 22 h to complete radical polymerization. Then, clozapine Fe₃O₄@SiO₂–MIPs nanoparticles were obtained. After the polymerization, the obtained polymers were washed with the mixture solution of methanol/acetic acid (8 : 2, v/v) using soxhlet extraction to remove the template molecules. The washing was continued until no CLZ or other compound was detected in supernatant. Finally, the product (Fe₃O₄@SiO₂–MIPs) was dried in vacuum at 70 °C. Fig. 2 presents the procedure for the preparation of Fe₃O₄@SiO₂–MIPs nanoparticles stepwise.

Fig. 2 Schematic representation of preparation for Fe₃O₄@SiO₂–MIPs nanoparticles.

Blank non-imprinted polymers, Fe₃O₄@SiO₂–NIPs (MNIPs), were prepared, in the absence of CLZ under the same condition as described above.

2.4. Absorption procedure equilibrium

For evaluation of the adsorption capacity of the MMIPs, the adsorption experiment was performed. 10 mg of MMIPs or MNIPs were equilibrated with 1 mL of concentrations (20 μg mL⁻¹ to 120 μg mL⁻¹) of CLZ dissolved in acetonitrile in a 2 mL centrifugal tube and shaken for 24 h at room temperature. Subsequently, the MMIPs and MNIPs were magnetically separated from the solution and supernatant was analyzed by HPLC. The amount of bound CLZ was calculated from the difference between initial and final concentrations in solution.

The equilibrium adsorption capacity S_b (μmol g⁻¹) was calculated as eqn (1):

In this equation, C_f (μg mL⁻¹) is the CLZ concentration of the supernatant solution after the adsorption, C_i (μg mL⁻¹) means the initial concentration of CLZ solution, V (mL) equals the volume of the initial CLZ solution, W (g) is the weight of the MMIPs or MNIPs and M (g mol⁻¹) is the molar mass of CLZ. Each test was carried out four times and mean ± SD was reported.

For investigation of the selectivity of the MMIPs, three different drugs (diazepam, fluoxetine and phenytoin) with CLZ were chosen and their structures are shown in Fig. 1. 20 mg of MMIPs or MNIPs were mixed with 1 mL of 20 μg mL⁻¹ solution of the studied compounds prepared in acetonitrile and then kept stirring for 24 h at room temperature. The particles were collected by a magnet. HPLC detections were applied in the supernatants of the studied solutions.

Interrelated adsorbed coefficients which including distribution coefficient and selectivity coefficient were evaluated.

The distribution coefficient was calculated according to eqn (2):

As the concentration of the adsorbed medium, C_a = (C_i − C_f) × V/W, and C_f is the free concentrations of the solution. In this equation, C_i (μg mL⁻¹), C_f (μg mL⁻¹), V (mL), and W (g) were described preciously.

Also, the selectivity coefficient (k) and relative selectivity coefficient (k′) were calculated by eqn (3) and (4) respectively:

In order to compare the selectivity of the MMIPs between CLZ and three different drugs, the selectivity coefficient (k) was calculated.

k_d1 stands for the distribution coefficient for CLZ and k_d2 stands for the distribution coefficients for three different drugs.

2.5. Extraction of CLZ from human serum samples

800 μL of ACN was added to 200 μL of serum in order to precipitate the serum proteins. After centrifugation (13 000 rpm for 10 min), 1.5 mL deionized water was added to 500 μL of supernatant. 40 mg of MMIPs were added into the solution and shook at room temperature for 30 min. A magnet was used to separate MMIPs from the solution followed by washing the MMIPs with 1 mL of methanol/acetic acid (9 : 1, v/v) for 20 min. 500 μL supernatants were evaporated to dryness and dissolved in 100 μL of methanol for further HPLC-UV analysis.

3. Results and discussion

3.1. Characterization of imprinted magnetic nanoparticles

Fourier transform infrared (FT-IR) spectra for magnetic Fe₃O₄ nanoparticles (MNPs), Fe₃O₄@SiO₂, Fe₃O₄@SiO₂–NH₂, Fe₃O₄@SiO₂–MIPs and Fe₃O₄@SiO₂–NIPs nanoparticles were recorded. Some of the important bands are given in Table 1. The bands centered around 3400 cm⁻¹ and 1620 cm⁻¹ are, respectively, assigned to the O–H stretching and deforming vibrations of adsorbed water. For all samples, the absorption peak at 570 cm⁻¹ is observed, corresponding to the Fe–O vibration from the magnetite phase.⁷⁰ The adsorption peaks at 1100 and 780 cm⁻¹ corresponds to the antisymmetric and symmetric stretching vibration of Si–O–Si bond in oxygen–silica tetrahedron, respectively.⁷⁰ Evidently, it indicates that the silica has been successfully coated on the surface of superparamagnetic Fe₃O₄ NPs. At FT-IR spectrum of Fe₃O₄@SiO₂–NH₂ nanoparticles; the peaks at 569, 1091, and 1591 cm⁻¹ are attributed to Fe–O (stretching vibration), Si–O–Si (asymmetric stretching) and N–H (bending), respectively. Also, the presence of several bands with medium intensity in 2861–2927 cm⁻¹ and 3238–3303 regions are allocated to C–H stretching of the propyl group and N–H stretching. The strong absorption bands at around 1727 and 1260 cm⁻¹, which were assigned to C=O stretching vibration of carboxyl (MAA) and C–O symmetric stretching vibration, respectively.⁷¹ Also, the peaks at 2930 cm⁻¹ and 2990 cm⁻¹ of Fe₃O₄@SiO₂–MIPs, indicated the presence of C–H stretching bands of both –CH₃ and –CH₂ groups. In addition, Fe₃O₄@SiO₂–MIPs and Fe₃O₄@SiO₂–NIPs showed almost the same characteristic absorption bands, indicating the complete removal of templates. These results indicate the co-polymerization of MAA and EGDMA on the surface Fe₃O₄@SiO₂–NH₂ in the presence of AIBN has been initiated.

Table 1 Infrared spectroscopic data for Fe₃O₄, Fe₃O₄@SiO₂, Fe₃O₄@SiO₂–NH₂, Fe₃O₄@SiO₂–MIPs and Fe₃O₄@SiO₂–NIPs nanoparticles

Vibration mode assignment	Fe3O4 (cm^−1)	Fe3O4@SiO2 (cm^−1)	Fe3O4@SiO2–NH2 (cm^−1)	Fe3O4@SiO2–MIPs (cm^−1)	Fe3O4@SiO2–NIPs (cm^−1)
Fe–O	570	571	569	570	572
Si–O–Si (symmetric)	—	789	771	776	782
Si–O–Si (asymmetric)	—	1097	1091	1109	1110
N–H	—	—	3100–3300	—	—
δ(CH2)	—	—	1472	1471	1469
C–H	—	—	2861–2927	2931–2985	2932–2989
C–O	—	—	—	1257	1260
C=O (ester)	—	—	—	1727	1728
O–H	3393	3401	3300–3400	3437	3428

---

Fig. 3 displays the XRD patterns of the Fe₃O₄ NPs, Fe₃O₄@–SiO₂ and Fe₃O₄@SiO₂–MIPs nanoparticles. As presented, six characteristic diffraction peaks (2θ = 30.1, 35.4, 43.1, 53.4, 57 and 62.6°), correspond to (220), (311), (400), (422), (511) and (440) reflections of inverse spinel Fe₃O₄ NPs, were also observed for Fe₃O₄@SiO₂ and Fe₃O₄@SiO₂–MIPs nanoparticles (reference JCPDS card no. 19-629) (Fig. 3).⁶⁸ This revealed that the surface modification of the Fe₃O₄ nanoparticles does not lead to their phase changes. It can be seen that the Fe₃O₄ obtained has highly crystalline cubic spinel structure which agrees with the standard Fe₃O₄ (cubic phase) XRD spectrum (PDF#88-0866). From Fig. 3b, we can see the XRD pattern of Fe₃O₄@SiO₂ showing an obvious diffusion peak at 2θ = 10–25°, generally considered as the diffusion peak of amorphous silica. For Fe₃O₄@SiO₂–MIPs nanoparticles, the broad peak was transferred to lower angles due to the synergetic effect of amorphous silica and polymer (Fig. 3c).

Fig. 3 XRD patterns of (a) Fe₃O₄, (b) Fe₃O₄@SiO₂ and (c) Fe₃O₄@SiO₂–MIPs nanoparticles; EDX spectra of (d) Fe₃O₄, (e) Fe₃O₄@SiO₂ and (f) Fe₃O₄@SiO₂–MIPs.

The broadening of each peak in XRD mean crystallite size was calculated by applying Scherrer's equation: D = Kλ/β cos α, where K is a constant (K = 0.9 for Cu-Kα), D is the average diameter in Å, β is the broadening of the diffraction line measured at half of its maximum intensity in radians, λ is the wavelength of the X-rays and α is the Bragg diffraction angle. According to the result calculated by Scherrer equation, it was found that the diameter of Fe₃O₄ nanoparticles obtained was about 12 nm and Fe₃O₄@SiO₂ microspheres were obtained with a diameter of about 20 nm due to the agglomeration of Fe₃O₄ inside nanospheres and surface growth of silica on the shell.⁷²

The components of the Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@SiO₂–MIPs nanoparticles were analyzed using energy dispersive spectroscopy (EDS) in Fig. 3d–f, respectively. In Fig. 3d, the EDS spectrum of Fe₃O₄ only showed peaks of iron and oxygen. As shown in the spectra of Fe₃O₄@SiO₂ in Fig. 3e, the presence of Si, O, and Fe signals indicates that the iron oxide particles are loaded into silica, and the higher intensity of the Si peak compared with the Fe peaks indicates that the Fe₃O₄ nanoparticles were trapped by SiO₂. Fig. 3f is the EDX spectrum of the Fe₃O₄@SiO₂–MIPs nanoparticles in which Fe, O, Si, C, and N are all present. This implied that MMIPs was coated on the surface of the Fe₃O₄@SiO₂ NPs.

The morphology and sizes of (a) Fe₃O₄ and (b) Fe₃O₄@SiO₂ particles were observed by transmission electron microscopy (TEM) as shown in Fig. 4.

Fig. 4 TEM images of (a) Fe₃O₄, (b) Fe₃O₄@SiO₂ and SEM images of (c) Fe₃O₄, (d) Fe₃O₄@SiO₂, (e) Fe₃O₄@SiO₂–MIPs and the size distributions of (f) Fe₃O₄, (g) Fe₃O₄@SiO₂ and (h) Fe₃O₄@SiO₂–MIPs, respectively.

The mesoporous silica shell on the surface of Fe₃O₄ is quite homogeneous and exhibits good monodispersity with estimated thickness of 8 nm (Fig. 4b). The morphology of Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@SiO₂–MIPs nanoparticles were also observed by SEM (Fig. 4c–e). The SEM images indicate the successful coating of the magnetic Fe₃O₄ particles. The Fe₃O₄@SiO₂–MIPs nanoparticles are approximately spherical shapes with a smooth surface morphology. The diameter of the nanoparticles is found to be approximately 100 nm and the thickness of MIP layer obtained was about 30 nm (Fig. 4e).

In this study, the hydrodynamic diameter of Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@SiO₂–MIPs nanoparticles is determined by the DLS technique (Fig. 4f–h). The average diameters of particles are evaluated to be about 12 nm for Fe₃O₄ (Fig. 4f), 20 nm for Fe₃O₄@SiO₂ (Fig. 4g) and 95 nm for Fe₃O₄@SiO₂–MIPs (Fig. 4h).

The histogram was proposed according to the results obtained from the XRD and TEM images.

The magnetic properties of the Fe₃O₄ NPs, Fe₃O₄@SiO₂ NPs and Fe₃O₄@SiO₂–MIPs NPs were indicated by magnetization hysteresis loops. Fig. 5 shows the absence of hysteresis phenomenon and indicates that product has superparamagnetism at room temperature.

Fig. 5 Magnetization curves at 300 K for (a) Fe₃O₄, (b) Fe₃O₄@SiO₂ and (c) Fe₃O₄@SiO₂–MIPs.

Magnetic measurement shows that pure Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@SiO₂–MIPs NPs have saturation magnetization values of 64.8, 40.3 and 27.3 emu g⁻¹, respectively.

These results indicated that the magnetization of Fe₃O₄ decreased considerably with the increase of amorphous silica and polymer. Nevertheless, these Fe₃O₄@SiO₂–MIPs NPs with superparamagnetic characteristics and high magnetization values can quickly respond to the external magnetic field and quickly redisperse once the external magnetic field is removed. This result reveals that the as-prepared Fe₃O₄@SiO₂–MIPs nanoparticle exhibit good magnetic responsible.

3.2. Adsorption isotherms

Scatchard analysis was shown that binding of CLZ to MMIP and its blank polymer (MNIP) was studied at different concentrations (Fig. 6). The data showed that CLZ binding to MMIP was significantly more than MNIP at all concentrations. At the beginning, both the amounts of CLZ bounded to MMIPs and MNIPs increased along the increment of initial concentrations until they reached the saturation level and then tended to be stable when the equilibrium concentration was equal or greater than 600 μM.

Fig. 6 (a) Adsorption isotherm of Fe₃O₄@SiO₂–MIPs and Fe₃O₄@SiO₂–NIPs using batch adsorption test (n = 6). Each point represents mean ± SD. Experiment conditions: 15 mg of polymer was incubated in 4 mL ACN with different concentrations of CLZ for 24 h at room temperature and (b) Scatchard plot to estimate the binding mechanism of Fe₃O₄@SiO₂–MIPs towards clozapine.

Two adsorption isotherms-mathematical models, including the Langmuir and Scatchard isotherm models were used to interpret the nature of the template–polymer binding affinity and to study further on the binding properties of Fe₃O₄@SiO₂–MIPs and Fe₃O₄@SiO₂–NIPs. The results were shown in Table 2. Interestingly, both the Langmuir isotherm and Scatchard isotherm show the high linear correlation coefficients (r_Langmuir = 0.9789, r_Scatchard = 0.9797).

Table 2 Langmuir and Scatchard isotherm constant for adsorption of CLZ on MMIPs

Isotherm	Bmax (μmol g^−1)	kd (Lm mol^−1)	r^2
Scatchard isotherm	12.78	14.5	0.9797
Langmuir isotherm	10.83	11.78	0.9789

---

In Scatchard isotherm:

where S_b is the amount of CLZ bound to the polymers at equilibrium (mg g⁻¹), C is the free CLZ concentration at equilibrium (μg mL⁻¹), k_d is the dissociation constant (mg L⁻¹), B_max is the apparent maximum binding amount (mg L⁻¹). Scatchard curves were obtained by taking S_b as the x-coordinate and S_b/C as the y-coordinate. From the Scatchard plot (Fig. 6b) one dissociation constants could be discerned, one was representing high affinity binding sites with a k_d of 14.5 μM, and B_max of 12.78 μmol g⁻¹ polymer.

3.3. Selectivity of the MMIP

The selectivity of Fe₃O₄@SiO₂–MIPs, for CLZ was evaluated in presence of other drugs in human serum. In this part, diazepam, fluoxetine, phenytoin and CLZ were selected as the potential interferes to investigate the selectivity of the imprinted nanoparticles based on their molecular weights and structures. The molecular recognition ability of MMIPs mainly depends on the binding ability which is closely related to the similarity between the template and the adsorbed molecules in functional groups, size and shape.

The specificity of Fe₃O₄@SiO₂–MIPs can be estimated by the imprinting factor of the selected clozapine between Fe₃O₄@SiO₂–MIPs and Fe₃O₄@SiO₂–NIPs. The imprinting factor α was determined according to eqn (6):

where S_b,MMIP and S_b,MNIP are the adsorption capacity of the same analyte Fe₃O₄@MIPs and Fe₃O₄@NIPs, respectively.

As shown in Fig. 7, the Fe₃O₄@SiO₂–MIPs revealed a significantly higher adsorption amount of clozapine than other drugs; however, the Fe₃O₄@SiO₂–NIPs did not show such a difference, indicating that the template molecule had a relatively higher affinity for the imprinted polymer than other drugs. Moreover, the α amount for clozapine was also much higher than other drugs, indicating that the Fe₃O₄@SiO₂–MIPs well recognized clozapine. The differences in their spatial structures and functional groups caused a mismatch in the holes and binding sites leading to less adsorption of other drugs and an imprinting factor (α) < 1. These results indicate the excellent imprinting efficiency of the present method.

Fig. 7 Percent of unbonded drug to MMIPs and MNIPs in optimum conditions.

3.4. Optimization of extraction conditions

For evaluation the use of MMIPs to separation of clozapine in human serum samples, we investigated the parameters affecting extraction performance, including time, amount of MMIPs and temperature. The adsorption of clozapine should achieve equilibrium to allow sufficient time to obtain satisfactory recovery. The extraction times were evaluated from 5 to 90 min. The results given in Fig. 8a show an increase in recovery from 5 to 8 min and above this time a slight decrease in recovery with increasing extraction time, although the differences were not significant (P > 0.05). This phenomenon may be due to re-dissolution of template or formation of covalent bound between CLZ and template. We selected 8 min for extraction time, which provided 89.35% ± 1.15% recovery.

Fig. 8 Optimization of extraction conditions (a) adsorption time, (b) amount of MMIPs and (c) temperature.

During extraction, the MMIPs were dispersed in the sample solution to rebind analyte, and we investigated the minimum amount of sorbent required for efficient recovery. MMIPs amounts from 10 to 100 mg were applied to extract clozapine from samples. The results indicated that 20 mg polymer was sufficient to yield 90.93% ± 2.53% recovery (Fig. 8b). Increasing the amount of MMIPs yielded no improvement in clozapine recovery. A slight decrease in recovery with increasing MMIPs amounts may be due to agglomeration of MMIPs nanoparticles and therefore decreasing of available site for adsorption of CLZ in MMIPs.

Extraction temperature was also investigated. We evaluated extraction temperature from 15 to 40 °C. The results given in Fig. 8c show a slight decrease in recovery with increasing extraction temperature above the 25 °C, although the differences were not significant (P > 0.05). We selected a 25 °C extraction, which provided 85.09% ± 0.78% recovery. A slight decrease in recovery with increasing temperature may be due to mobility of template. The mobility is higher at higher temperature.

3.5. Precision and accuracy

Intra-day and the inter-day precision and accuracy were determined by assaying blank plasma spiked with five different concentrations of clozapine. Intra-day precision was assessed by assaying five samples at three drug concentrations (50, 200 and 1000 ng mL⁻¹). Inter-day precision was evaluated by assaying (seven days) ten samples at two concentrations (100 and 500 ng mL⁻¹). In the range investigated intra-day and inter-day coefficients of variation (C.V.) were less than 6 and 4% for clozapine (Table 3).

Table 3 Intra-day and inter-day precision and accuracy of the method for clozapine

Spiked value (ng ml^−1)	Measured value (ng ml^−1) (mean ± S.D.)	Intra-day (n = 5) C.V. (%)	Inter-day (n = 10) C.V. (%)	Relative error (%)
50	49.56 ± 4.1	6.1		2.1
100	88.43 ± 3.7		3.4	3
200	198.13 ± 6.4	3.2		3.7
500	490.78 ± 13.67		3.6	2
1000	999.57 ± 27.02	3		0.08

---

3.6. Reusability

Repeated binding/removal experiments were performed to test the stability and reusability of Fe₃O₄@SiO₂–MIPs. The recovery of CLZ on Fe₃O₄@SiO₂–MIPs decreased 6.67% after eight cycles, suggesting they can be recycled.

4. Conclusions

In this study, Fe₃O₄@SiO₂–MIPs was synthesized with a MIPs layer on Fe₃O₄@SiO₂–NH₂ NPs with a uniform core–shell structure by surface imprinting and nanotechniques for adsorbing and recognizing of CLZ from human serum. The obtained Fe₃O₄@SiO₂–MIPs was characterized via FT-IR, XRD, TEM, SEM, DLS and VSM. The selectivity recognition properties of the Fe₃O₄@SiO₂–MIPs were evaluated and the results showed that the Fe₃O₄@SiO₂–MIPs had high adsorption capacity and selectivity for CLZ. Selectivity of the Fe₃O₄@SiO₂–MIPs procedure was investigated using CLZ and some drugs that could be present, simultaneously, in serum of patients. The prepared Fe₃O₄@SiO₂–MIPs exhibited excellently specific recognition and saturation magnetization. Moreover, it could be easily recovered by external magnetic field, leading to a fast and selective recognition of CLZ from human serum. After Fe₃O₄@SiO₂–MIPs were reused and regenerated eight times, the adsorption capacity was still excellent. The approach reported here could provide a selective, cost-efficient and sensitive method for serum CLZ level determination in therapeutic range.

Acknowledgements

The authors are grateful to the council of Iran National Science Foundation and University of Shiraz for their unending effort to provide financial support to undertake this work.

References

1. W. Yang, F. Jiao, L. Zhou, X. Chen and X. Jiang, Appl. Surf. Sci., 2013, 284, 692.
2. W. Liu, H. Zhao, Y. Yang, X. Liu and B. Xu, Appl. Surf. Sci., 2013, 277, 146.
3. R. Jackson, I. Petrikovics, E. P. C. Lai and J. C. C. Yu, Anal. Methods, 2010, 2, 552.
4. S. Chen, J. Chu and X. Li, Appl. Surf. Sci., 2013, 284, 745.
5. X. L. Sun, X. W. He, Y. K. Zhang and L. X. Chen, Talanta, 2009, 79, 926.
6. T. Yamazaki, S. Ohta, Y. Yanai and K. Sode, Anal. Lett., 2003, 36(1), 75.
7. X. Su, X. Li, J. Li, M. Liu, F. Lei, X. Tan, P. Li and W. Luo, Food Chem., 2015, 171, 292.
8. C. Schirmer and H. Meisel, Anal. Bioanal. Chem., 2008, 392, 223.
9. G. Wulff, Chem. Rev., 2002, 102, 1.
10. A. L. Jenkins, R. Yin and J. L. Jensen, Analyst, 2001, 126, 798.
11. A. Molinelli, R. Weiss, B. Mizaikoff and J. Agric, Food Chem., 2002, 50, 1804.
12. S. Scorrano, L. Longo and G. Vasapollo, Anal. Chim. Acta, 2010, 659, 167.
13. F. F. Chen, R. Wang and Y. P. Shi, Talanta, 2012, 89, 505.
14. F. F. Chen, G. Y. Wang and Y. P. Shi, J. Sep. Sci., 2011, 34, 2602.
15. N. Masque, R. M. Marce and F. Borrull, Trends Anal. Chem., 2001, 20, 477.
16. B. Sellergren, J. Chromatogr. A, 2001, 906, 227.
17. K. Kim and D. Kim, J. Appl. Polym. Sci., 2005, 96, 200.
18. N. Masque, R. M. Marce, F. Borrull, P. A. G. Cormack and D. C. Sherrington, Anal. Chem., 2000, 72, 4122.
19. L. L. Zhu and X. J. Xu, J. Chromatogr. A, 2003, 991, 151.
20. S. W. Chuang, J. Rick and T. C. Chou, Biosens. Bioelectron., 2009, 24, 3170.
21. B. C. Karlsson, A. M. Rosengren, I. Näslund, P. O. Andersson and I. A. Nicholls, J. Med. Chem., 2010, 53, 7932.
22. J. He, R. Lv, J. Zhu and K. Lu, Anal. Chim. Acta, 2010, 66, 1215.
23. J. Bastide, J. P. Cambon, F. Breton, S. A. Piletsky and R. Rouillon, Anal. Chim. Acta, 2005, 542, 97.
24. J. L. Urraca, M. C. Moreno-Bondi, G. Orellana, B. Sellergren and A. J. Hall, Anal. Chem., 2007, 79, 4915.
25. Y. P. Duan, C. M. Dai, Y. L. Zhang and L. Chen, Anal. Chim. Acta, 2013, 758, 93.
26. A. Prieto, S. Schrader, C. Bauer and M. Möder, Anal. Chim. Acta, 2011, 685, 146.
27. Z. Li, M. Day, J. Ding and K. Faid, Macromolecules, 2005, 38, 2620.
28. T. Alizadeh, M. R. Ganjali, M. Zare and P. Norouzi, Food Chem., 2012, 130, 1108.
29. T. A. Sergeyeva, O. A. Slinchenko, L. A. Gorbach, V. F. Matyushov, O. O. Brovko, S. A. Piletsky, L. M. Sergeeva and G. V. Elska, Anal. Chim. Acta, 2010, 659, 274.
30. Y. Watabe, K. Hosoya, N. Tanaka, T. Kubo, T. Kondo and M. Morita, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 2048.
31. H. Sanbe and J. Haginaka, J. Pharm. Biomed. Anal., 2003, 30, 1835.
32. H. Sambe, K. Hoshina, K. Hosoya and J. Haginaka, J. Chromatogr. A, 2006, 1134, 16.
33. F. Tan, D. Sun, J. Gao, Q. Zhao, X. Wang, F. Teng, X. Quan and J. Chen, J. Hazard. Mater., 2013, 244–245, 750.
34. G. Pan, Y. Zhang, X. Guo, C. Li and H. Zhang, Biosens. Bioelectron., 2010, 26, 976.
35. W. Chi, H. Shi, W. Shi, Y. Guo and T. Guo, J. Hazard. Mater., 2012, 227–228, 243.
36. L. Li, X. He, L. Chen and Y. Zhang, Chem.–Asian J., 2009, 4, 286.
37. L. Xu, J. Pan, J. Dai, X. Li, H. Hang, Z. Cao and Y. Yan, J. Hazard. Mater., 2012, 233–234, 48.
38. X. Liu, D. Yu, Y. Yu and S. Ji, Appl. Surf. Sci., 2014, 320, 138.
39. Y. Hiratsuka, N. Funaya, H. Matsunaga and J. Haginaka, J. Pharm. Biomed. Anal., 2013, 75, 180.
40. J. Pan, L. C. Xu, J. Dai, X. Li, H. Hang, P. Huo, C. Li and Y. Yan, Chem. Eng. J., 2011, 174, 68.
41. S. Abramson, W. Safraou, B. Malezieux, V. Dupuis, S. Borensztajn, E. Briot and A. Bée, J. Colloid Interface Sci., 2011, 364, 324.
42. M. F. Shao, F. Y. Ning, J. W. Zhao, M. Wei, D. G. Evans and X. Duan, J. Am. Chem. Soc., 2012, 134, 1071.
43. H. X. Wu, L. H. Tang, L. An, X. Wang, H. Q. Zhang, J. L. Shi and S. P. Yang, React. Funct. Polym., 2012, 72, 329.
44. J. Liu, B. Wang, S. B. Hartono, T. T. Liu, P. Kantharidis, A. P. J. Middelberg, G. Q. Lu, L. Z. He and S. Z. Qiao, Biomaterials, 2012, 33, 970.
45. A. H. Lu, E. L. Salabas and F. Schuth, Angew. Chem., Int. Ed., 2007, 46, 1222.
46. F. H. Chen, Q. Gao and J. Z. Ni, Nanotechnology, 2008, 19, 165103.
47. S. H. Xuan, S. F. Lee, J. T. F. Lau, X. M. Zhu, Y. X. J. Wang and F. Wang, ACS Appl. Mater. Interfaces, 2012, 4, 2033.
48. J. Zhang, W. Sun, L. Bergman, J. M. Rosenholm, M. Lindén, G. Wu, H. Xu and H. C. Gu, Mater. Lett., 2012, 67, 379.
49. Y. H. Deng, D. W. Qi, C. H. Deng, X. M. Zhang and D. Y. Zhao, J. Am. Chem. Soc., 2008, 130, 28.
50. B. Liu, W. X. Xie, D. P. Wang, W. H. Huang, M. J. Yu and A. H. Yao, Mater. Lett., 2008, 62, 3014.
51. Y. H. Deng, Y. Cai, Z. K. Sun, J. Liu, C. Liu, J. Wei, W. Li, Y. Wang and D. Y. Zhao, J. Am. Chem. Soc., 2010, 132, 8466.
52. B. Panella, A. Vargas and A. Baiker, J. Catal., 2009, 261, 88–93.
53. F. F. Chen, X. Y. Xie and Y. P. Shi, J. Chromatogr. A, 2013, 1300, 112.
54. C. Hu, J. Deng, Y. Zhao, L. Xia, K. Huang, S. Ju and N. Xiao, Food Chem., 2014, 158, 366.
55. X. Kong, R. Gao, X. He, L. Chen and Y. Zhang, J. Chromatogr. A, 2012, 1245, 8.
56. T. Sharma, C. Hughes, W. Soni and V. Kumari, Psychopharmacology, 2003, 169, 398.
57. J. M. Alvir, J. A. Lieberman, A. Z. Safferman, J. L. Schwimmer and J. A. Schaff, N. Engl. J. Med., 1993, 329, 162.
58. W. R. Jin, Q. Xu and W. Li, Electrophoresis, 2000, 21, 1415.
59. Y. L. Shen, H. L. Wu, W. K. Ko and S. M. Wu, Anal. Chim. Acta, 2002, 460, 201.
60. C. S. P. Sastry, T. V. Rekha and A. Satyanarayana, Microchim. Acta, 1998, 128, 201.
61. M. Kollroser and C. Schober, Rapid Commun. Mass Spectrom., 2002, 16, 1266.
62. M. Esmaeilpour, J. Javidi, F. Dehghani and F. Nowroozi Dodeji, RSC Adv., 2015, 5, 26625.
63. J. Jaber and E. Mohsen, Colloids Surf., B, 2013, 102, 265.
64. M. Esmaeilpour, A. R. Sardarian and J. Javidi, J. Organomet. Chem., 2014, 749, 233.
65. M. Esmaeilpour, J. Javidi and M. Zandi, Mater. Res. Bull., 2014, 55, 78.
66. J. Javidi, M. Esmaeilpour and F. Nowroozi Dodeji, RSC Adv., 2015, 5, 308.
67. M. Esmaeilpour, J. Javidi, F. Dehghani and F. Nowroozi Dodeji, New J. Chem., 2014, 38, 5453.
68. M. Esmaeilpour, J. Javidi, F. Nowroozi Dodeji and M. Mokhtari Abarghoui, J. Mol. Catal. A: Chem., 2014, 393, 18.
69. M. Esmaeilpour, J. Javidi, F. Nowroozi Dodeji and M. Mokhtari Abarghoui, Transition Met. Chem., 2014, 39, 797.
70. (a) C. Yang, J. Wu and Y. Hou, Chem. Commun., 2011, 47, 5130; (b) M. Salavati-Niasari, J. Javidi and M. Dadkhah, Comb. Chem. High Throughput Screening, 2013, 16, 458.
71. K. Yoshimatsu, K. Reimhult, A. Krozer, K. Mosbach, K. Sode and L. Ye, Anal. Chim. Acta, 2007, 584, 112.
72. M. Esmaeilpour, A. R. Sardarian and J. Javidi, Appl. Catal., A, 2012, 445–446, 359.

---