Preparation of Magnetic Molecularly Imprinted Polymers for Selective Recognition and Determination of Clenbuterol in Pork Samples

Magnetic molecularly imprinted polymer (MMIP) was successfully synthesized with acrylamide as a functional monomer and clenbuterol (CLB) as a template molecule. -e synthesized MMIPs were characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD), and Fourier transform infrared spectrometry (FT-IR). MMIPs were used to identify and bind CLB as a solid phase extraction material. -e experiment data were fitted by the Freundlich isotherm adsorption model. -e results show that MMIPs have excellent recognition performance for CLB. MMIPs were successfully applied as adsorbents to preconcentrated CLB in pork samples and detected by HPLC with UV. -e limit of detection (LOD) and limit of quantification (LOQ) were 4.27 μg/L and 14.2 μg/L, respectively. -e spiked recovery rates ranged from 94.44% to 102.29%. -erefore, the prepared MMIPs can be used for selective preconcentration of CLB content in complex animal-derived food samples.


Introduction
Clenbuterol (CLB) is one of the most common β-adrenergic agonists that can be used in the treatment of diseases such as bronchial asthma and chronic bronchitis as a bronchodilator agent in the clinic treatment [1][2][3]. Moreover, a large amount of CLB can be used as feed additives to significantly promote the growth of the animal and increase the lean rate. It promotes the synthesis of proteins in muscles, especially skeletal muscles, and inhibits the synthesis of fat by changing the metabolic pathways in animals, thereby accelerating the growth rate, increasing the lean meat and improving the carcass quality. erefore, CLB is also known as a lean meat agent [4][5][6].
As a feed additive, CLB can be used to increase the lean rate by using a large dose, which is usually more than 10 times that of human medicine. However, a large dosage, long use time, and slow metabolism will lead to a problem which is undesirable drug residues in animal-originated foodstuffs [7]. Long-term ingesting CLB from animal foods can cause abnormal physiological reactions such as dizziness, vomiting, muscle tremors, and even death [8]. In addition, illegal and uncontrolled use of CLB can also increase the environmental burden and cause environmental pollution [9]. e Chinese government clearly stipulates that the use of clenbuterol in pig breeding is prohibited. According to the Chinese government standard, clenbuterol should be no more than 0.01 mg kg −1 . Although many countries and regions have forbidden the use of CLB in livestock production, it has been still added in the fodder for animals to seek benefits [10][11][12]. erefore, it is urgent to establish a simple, rapid, and efficient method to detect the residue of CLB to ensure food safety.
Currently, there have been several reports on methods for determining residue of CLB, such as biosensors, immunology, and chromatography [13]. However, all of the above methods have more or fewer shortcomings. e limitation of biosensors [14] is the lack of stability and reproducibility. Immunological methods, such as flow immunoassay [15], immunochromatographic assay (ICA) [16], enzyme-linked immunosorbent assay (ELISA) [17], and time-resolved fluoroimmunoassay (TRFIA) [18], are less sensitive while the preparation, stability, and storage of antibodies always imply high costs [13]. Although the sample could be separated by chromatography and detected accurately by MS, time-consuming pretreatment, complex operation, and the high cost of the whole experiment are often cited as problems. Also, chromatography-mass spectrometry suffers from drawbacks of matrix effects which affect the selectivity, detection limits, maintenance frequency, and quantitative aspects. e most important shortcoming of these procedures is the lack of selectivity for target molecules. Moreover, the pretreatment of samples by these methods are mostly complicated.
In order to avoid the above weakness and promote advantages, researchers have been trying to establish new methods. Recently, molecular imprinting technology (MIT) has received more and more attention and is used in many fields, especially in solid phase extraction [19,20]. It is a technique for adsorbing target molecules by preparing molecularly imprinted polymers (MIPs) with specificity and selectivity [21].
MIP is a three-dimensional network polymer with specific recognition and selective adsorption, which is usually synthesized by free-radical polymerization. First, the template molecule interacts with the functional monomer to form tailor-made binding sites; and then crosslinkers are distributed around the template molecules, which polymerize the highly crosslinked polymers. After the crosslinking reaction, the template molecules are eluted with a special solvent, the binding cavities which are capable of specifically recognizing the target molecules are revealed in the crosslinked material [22][23][24][25][26].
Alternatively, magnetic separation technology has also received a lot of attention due to its being fast, economical, and efficient [27][28][29]. When the magnetic separation technology and molecular imprinting technology are combined, the magnetic molecularly imprinted polymers (MMIPs) are formed. ese MMIPs not only show excellent specific selective binding for target analytes but also can be quickly separated and easily recovered from the samples by an external magnet without centrifugation or filtration.
In this work, the MMIPs were synthesized and characterized as selective adsorbents for CLB. e Fe 3 O 4 nanoparticles are encapsulated in the polymer to give it magnetic properties. e kinetic studies and adsorption capacity on CLB adsorption by MMIPs were evaluated in detail. Finally, CLB was successfully preconcentrated by prepared MMIPs and detected with HPLC-UV in actual pork samples. (Darmstadt, Germany). All other organic solvents and inorganic reagents were of analytical reagent grade and provided by local suppliers. Pock was purchased from the local supermarket (Xi'an, China). All HPLC solutions were filtered using a filter (HA-0.45 μm) before being injected into the HPLC system.

Preparation of Fe 3 O 4 Magnetic Particles. Fe 3 O 4 magnetic
particles were synthesized by improved chemical coprecipitation method according to the previous study [30,31]. Briefly, FeCl 3 •6H 2 O (5.11 g) and FeCl 2 •4H 2 O (1.83 g) were dissolved in 80 mL of deoxygenated water in a 250 mL fournecked flask, and the mixture was constantly stirred under nitrogen protection. en 60 mL of ammonium hydroxide solution (5%) was added drop by drop when the temperature was raised to 80°C. e mixture was stirred vigorously for 60 min at 80°C. Subsequently, the Fe 3 O 4 nanoparticles were collected by using an external magnet and washed with deionized water several times until the pH � 7.
e PEG-modified Fe 3 O 4 was prepared according to previous work with some minor modifications. e prepared Fe 3 O 4 (2.0 g) and PEG (10.0 g) were dissolved in deoxygenated water (30 mL) by stirring for 30 min and sonicating for 20 min to obtain the homogeneously dispersed solution, which was stored in a dark place for later use.

Preparation of the MMIPs.
e MMIPs were prepared as follows: firstly, an appropriate amount of clenbuterol (0.5 mmol), as the template molecule, and AM (4 mmol), as the functional monomer, were added to a glass flask containing 25 mL acetonitrile. en, the mixture was stored in the dark for 18 h at room temperature to obtain the prepolymerization solution. After that, the prepolymerization solution and PEG-Fe 3 O 4 particles were dispersed in 80 mL of doubly distilled water and copolymer monomer (St, 87.4 mmol), crosslinker (EDGMA, 30 mmol), and initiator (AIBN, 0.6 mmol) were added to the mixture in a 250 mL three-neck flask.
e processed mixture was placed in a water bath and stirred for the polymerization at 70°C under nitrogen protection for 12 h. After polymerization, MMIPs were collected by an extra magnetic field and washed by a mixture of methanol-acetic acid (9 : 1, v/v) to remove the templates, and then rinsed with methanol. e synthesized MMIP (30 mg) was added to methanol (1 mL). After shaking for 15 min, HPLC was used to detect whether the template molecules were removed. Finally, the MMIPs were dried in a vacuum. For comparison, the nonimprinted polymers (NIPs) were also prepared as the same method mentioned above in the absence of the template.

Adsorption Capacity Experiment.
To evaluate adsorption properties of MMIPs for clenbuterol, the batch adsorption test was performed as follows: 30 mg MMIPs or MNIPs were immersed into the 2 mL centrifuge tube containing methanol solution of clenbuterol (100-300 μg L −1 , 1 mL) followed by shaking for 15 min at room temperature. MMIPs or MNIPs were separated from the solution by using a magnet. e concentration of clenbuterol in the supernatant was analyzed by HPLC-UV after filtration through a 0.45 μm microporous membrane. In order to evaluate the binding parameters of MMIP and MNIP, the Freundlich isotherm (FI) model was further used to further process the data from the adsorption experiments.
In adsorption kinetics experiments, 30 mg of MMIPs and MNIPs were added to 1 mL of clenbuterol methanol solution (160 μg L −1 ). e mixture was shaken mechanically for 5, 10, 15, 20, 30, 45, and 60 min in a shaken bed at room temperature to choose the optimal time of competitive adsorption.

Determination of CLB in Pork Samples.
One hundred grams of pork (obtained from a local market) was broken with a tissue shredder in 100 mL pure water. en MMIPs (30 mg) were added to the real sample homogenate solution (1 mL), followed by being shaken at 25°C for 15 min. And then, MMIPs were separated from the solution using an external magnet. e CLB was eluted from the MMIPs using 1 mL of methanol-acetic acid (9 : 1, v/v) solution. e eluted solution was filtered by a membrane filter (0.25 μm), dried by nitrogen flow, and dissolved in 0.1 mL of methanol for the detection by HPLC-UV analysis. Figure 1 shows that the preparation process of MMIPs is mainly divided into two steps. e first is the synthesis of Fe 3 O 4 nanoparticles. At present, the chemical coprecipitation method is widely used to prepare magnetic nanoparticles because of its simple operation and good reproducibility. erefore, Fe 3 O 4 nanoparticles are prepared by this method with superparamagnetism in this experiment. However, the prepared Fe 3 O 4 nanoparticles are prone to magnetic flocculation, so that the result is that nanoparticles lack monodispersity.

Preparation and Characterization of MMIPs.
erefore, the surface of Fe 3 O 4 particles was modified with PEG to avoid electrostatic agglomeration.
In noncovalent molecular imprinting, the interaction between functional monomers and template molecules affects the recognition performance of MMIPs directly. erefore, acrylamide (AM) and methacrylic acid (MAA) were selected as functional monomers to evaluate optimal one. In theory, both AM and MAA could produce hydrogen bonds with clenbuterol, but the experimental results show that the specific recognition ability of MMIPs synthesized with AM as a functional monomer is stronger. Further, styrene was used as a comonomer in this experiment because the unsaturated bond in its structure contributes to the formation of a crosslinked structure and increases the structural stability of the polymers.
SEM was used to measure the morphological features of MMIPs and MNIPs in Figure 2. As can be seen from the figure, the surface of the MMIPs is rough and porous, while the surface of the MNIPs is relatively regular. As a result, the porous surface and imprinting sites of the MMIPs are more conducive to specific adsorption than MNIPs. FT-IR was used to analyze the main functional groups of synthesized results. Figure 4 shows the infrared spectrum of Fe 3 O 4 nanoparticles, MMIPs, and MNIPs. e typical bands at 3028 and 2933 cm −1 can be attributed to the C-H aromatic stretching vibrations of styrene units. e absorption peak of C � O at 1723.9 cm −1 is due to the crosslinking agent EGDMA reacted with AM. Compared with NIP, weak absorption peaks appeared at 1632.8 and 1599.8 cm −1 , which attributed to carbonyl and bending vibration of the imino group in CONH 2 . It indicated that most AM and crosslinking agent had crosslinking polymerized during the preparation of MIPs, and only some AM residues were found. Compared to curve 3 (black), the band of Fe-O at 547.1 cm −1 was observed in MMIPs, which confirmed that Fe 3 O 4 was implanted into the polymer network successfully.

Adsorption and Analytical Kinetic
Evaluation. An adsorption kinetics experiment was performed on MMIPs to select the best extraction time. Figure 5(a) shows that the amount of adsorption increased sharply in the first 15 min. e adsorption equilibrium was reached at 15 min, and the maximum adsorption amount was obtained, so 15 min was selected as the extraction time. Figure 5(b) shows the desorption kinetics of CLB by MMIP, and the desorption time was optimized. After 15 min, the desorption ratio is close to 100%. After desorption, MMIPs can be quickly separated and recovered from the sample solution by an external magnetic field.

Adsorption Isotherm Equation.
To assess the binding energy, binding site type and distribution between MMIPs or MNIPs and imprinted molecules, adsorption isotherms were adopted using different concentrations of CLB. Figure 6(a) shows the adsorption isotherms of MMIPs and    MNIPs. When the adsorption reaches equilibrium, there is a certain relationship between the adsorption amount B and the residual amount F. It is possible to observe an increase of adsorption amount of polymer materials with increasing of the initial concentration of CLB, and the adsorption capacity of MMIPs for CLB is greater than that of MNIPs. Some studies have shown that the Freundlich equation has good applicability to noncovalently imprinted materials [30,32]. erefore, the Freundlich isothermal (FI) adsorption model was adopted to analyze the data with adsorption isotherms. e FI equation is expressed as follows [32]: where B and F are the concentrations of bound and free analyte, respectively; a and m are characteristic parameters of the Freundlich isotherm equation: a is the total number of binding sites, which represents the adsorption capacity of the adsorbent; m is the heterogeneity index with a value from 0-1. e value of m closes to 0 or 1 indicates that the adsorption sites are heterogeneous and homogeneous, respectively. And the value of a and m can be obtained from the following equation: e number of binding sites with given affinity (N(k)) of a polymer can be calculated by equation (3), where K is the affinity constant, K � 1/F.
e binding site N k min−k max and the apparent average binding constant K k min−k max per gram of polymer can be calculated from equation (4) and equation (5), respectively. N k min−k max is the number of binding sites per gram of polymer. K k min−k max is apparent average binding constant. ey can be calculated from equation (4) and equation (5), respectively, where K min � 1/F max ; K max � 1/F min . F max and F min are the maximum and minimum solute concentrations of the liquid phase at adsorption equilibrium, respectively; a and m are the characteristic parameters of the Freundlich isothermal equation.

Journal of Chemistry
e FI affinity distribution model was further used to analyze the adsorption data. e affinity distribution curve ( Figure 6(b)) was obtained by plotting the value of N(k) and log K, indicating the number of binding sites with binding ability. e value of log K represents the binding energy as the abscissa. It can be seen from the figure that the number of binding sites of MMIPs is always higher than that of MNIPs when given the binding energy within the experimental concentration range. It indicates that there are specific adsorption sites in the MMIPs material, and the imprinting effect is evident. e results are exhibited in Table 1. e total number of binding sites and binding constants of MMIPs is higher than MNIPs. e binding sites N, K of MMIPs per gram of polymer are larger than MNIPs. ese results indicate that the target molecule plays an important role in the formation of specific sites during the imprinting process.

Method Validation.
In order to verify the accuracy of the method, the linear relationship, correlation coefficient (r), the limit of detection (LOD), the limit of quantification (LOQ), and spiked recovery experiment were investigated.
By drawing the standard curve, the linear regression equation is y � 7.724x + 278 (r � 0.9694) (Figure 7). e values of LOD and LOQ were measured to be 3 and 10 times the signal-to-noise ratio (S/N), respectively. eir values were 4.27 and 14.2 μg L −1 .
At the same time, the spiked sample recovery experiment was used to assess the reproducibility and accuracy of the method.
ree different concentrations of samples were added to the CLB samples, and the recovery was measured and calculated. e concentration of the base solution is 0.25 μmol L −1 . e recovery values were calculated by the formula below: e results were shown in Table 2. e recoveries ranged from 94.44 to 102.29%, and the relative standard deviation (RSD) ranged from 3.48 to 7.14%. e same sample was measured 5 times and the RSD was calculated to be 4.93%.
3.5. Analysis of Real Samples. In order to investigate the extraction and enrichment ability of MMIPs on actual    Figure 8(a) shows that CLB cannot be detected directly because the concentration is low without enrichment. Figure 8(b) also shows the sample could not be extracted by MNIPs, which proves that its enrichment is nonspecific adsorption. Figure 8(c) shows that a higher peak appeared and highly selective adsorption was achieved after the sample was extracted by MMIPs. e content of CLB in pork was calculated resulting in 9.87 ± 0.48 μg kg −1 . e result showed that MMIPs could be used to enrich and detect CLB in actual samples.     Journal of Chemistry

Conclusions
In summary, this study synthesized MMIPs of CLB for the first time. e synthesized MMIPs were characterized by SEM, FT-IR, and XRD. Adequate Fe 3 O 4 nanoparticles are coated in the MMIPs, which ensures that the MMIPs can be easily and quickly separated from the complex solution through an external magnetic field without centrifugation or filtration. rough isothermal adsorption experiments, it was confirmed that MMIPs have good specific recognition and selective adsorption ability. e high recovery of spiked recovery experiments demonstrates that this approach is effective. erefore, the method can be used for the residue detection of clenbuterol in animal products such as pork or pig urine.

Data Availability
e authors confirm that all data underlying the findings of this study are fully available without restriction.

Conflicts of Interest
e authors declare no conflicts of interest.