Lu Yunkai, Zhao Ning, Qin Xinying, Liu Yue, Lu Guodong
(College of Chemistry and Environmental Science, Hebei University, Key Laboratoryo f Analytical Science and Technology of Hebei Province, Baoding 071002, China)

Abstract A new molecularly imprinted polymer was synthesised based on sacrificial mesoporous silica (Si-MIP) for the solid-phase extraction of fluoroquinolone residues in edible animal products. The Si-MIP was prepared by filling mesoporous silica with the template molecule and monomers followed by polymerisation, then the silica particles materials were removed from the obtained imprinted composites. The binding capacity was evaluated by static adsorption and Scatchard analysis, which showed that the dissociation constant () and the maximum binding capacity (_max) were 178.9 mg L^-1 and 73.31 mg g^-1 for high affinity binding site, and 45.54 mg L^-1 and 40.82 mg g^-1 for lower affinity binding site, respectively. The largest selectivity coefficient for ofloxacin in the presence of enrofloxacin was found to be 3.79, the largest relative selectivity coefficient between ofloxacin and enrofloxacin over 3, along with a fast kinetics comparing with MIP without silica template. The SPE procedure was optimized, and the accuracy and selectivity of Si-MIP cartridge process developed were verified using a MIP cartridge and a classical C18 cartridge as the SPE matrix during control experiments. The quantification and detection limits in tissue samples of ofloxacin, enrofloxacin and flumequin and were established at 18 m g kg^-1, 21 m g kg^-1 and 42 m g kg^-1, respectively.

 

1. INTRODUCTION
Molecular imprinting technique (MIT) is an increasingly developing technique for preparing polymers with desired and predetermined selectivity, and provides specific binding sites or catalytic sites in molecularly imprinted polymers (MIPs). MIPs had been attracted extensive attention and have been used extensively in sensors, immunoassay-type binding assays in place of antibodies and the separation techniques, which involve affinity chromatography, capillary electrochromatography, solid phase extraction (SPE), thin layer chromatography and membrane separation^[1]. A particularly promising application of MIPs is as selective sorbents in SPE for the cleanup and preconcentration of compounds from low concentrations or complex matrices^{[2, 3]}. The key of molecularly imprinted solid phase extraction (MISPE) development is preparations of the molecular imprinting sorbents.
The conventional method is to synthesize the MIPs in bulk thermal-(or photo-)polymerisation that produces a monolithic polymer that has to be grinded and sieved, resulting in irregularly shaped materials with heterogeneous size and porosity. Although this technique has led to highly selective materials for a multitude of analytes, such particles are not well suited as packing materials for SPE. For example, high backpressures and low mass transfer kinetics are usually observed. Sacrificial materials can overcome these problems, and these synthesic methods have been published, such as filling support particles^[4]
In this study, we apply a MIPs preparation protocol based on sacrificial mesoporous silica^[4]to prepare a new SPE sorbent with high selectivity and high mass transfer kinetics for selective separation and preconcentration of the veterinary residue (e.g. quinolones) from edible animal products.
Quinolones are antibacterial agents widely used in the treatment of infections in both humans and animal. Their primary target is the bacterial enzyme DNA gyrase or topoisomerase II, which renders the DNA molecule compact and biologically active^[5]. Quinolones are widely used in veterinary medicine, and the problem of veterinary residue at animality foodstuff become societal focus. The European Union and the Joint FAO/WHO Expert Committee on Food Additives (JECFA) have established maximum residue limits (MRL) for several quinolones. The harmful of veterinary residue for people and environment is chronic, long-dated and cumulate. When prople eat the animality foodstuff which go beyond standard mete of the veterinary residue, the health or life will be endanger^[6]
Quinolone residue analysis involves extraction with an appropriate solvent followed by one or more clean-up processes and determination by HPLC, LC-MS, CE, CE-MS, ELISA. The sample treatment is important procedure before quinolone residue determination in order to reduce the matrix interference and enrich the analytes. Selective extraction of quinolone drug from foodstuff is usually achieved by solvent extraction and solid phase extraction (SPE), the latter is more rapid, simple, economical, environmental friendly and easy automatization than the traditional liquid Cliquid extraction (LLE). Increasing the selectivity of sorbent in the extraction of analytes and developing new efficient cleanup techniques are highly attractive for monitoring trace analytes in complex samples, So, development of new solid sorbents for selective separation quinolone residues in food, therefore, is of great significance^[7]
The development of the molecular imprinting technique about quinolone medicament have some study and put up huge potential and advantage^[8-12]. In this work, A new preparation method of molecularly imprinted polymer based on sacrificial mesoporous silica (Si-MIP) was investigated, and the quinolone imprinted polymer was used as SPE sorbent for separation and preconcentration of the quinolone residue from edible animal products.

2. EXPERIMENTAL
2.1 Chemicals
All the chemicals used were of analytical reagents: Methacrylic acid (MAA) and ethylene glycol dimethacrylate (EGDMA) from Sigma-Aldrich (Shanghai) Trading Co., Ltd. (Shanghai, China) and cleaned to remove the inhibitor prior to polymerization. 2,2-Azobisisobutyronitrile (AIBN)from Beijing Chemical Reagent company (Beijing, China) recrystallized prior to use. Ofloxacin (OFL), enrofloxacin (ENR) and flumequine (FLU) were obtained from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). The structures of the studied compounds are shown in Fig. 1. Anhydrous alcohol, acetone, acetonitrile (ACN), chloroform and methanol were all of HPLC grade and purchased from Tianjing Kermel Chemical Reagents Development Centre (Tianjin, China). Trichloroacetic acid (TCA) and acetic acid (analytical grade) was purchased from Jinli Industries Co. (Tianjin, China). All of the solutions were filtered through a 0.45 m m membrane filter (Millipore) before use. Doubly deionized water (DDW) was used throughout.
Commercially available silica particles (20 m m mean diameter, 0.7 ml g^-1 pore volume, 8.0 nm pore size) were obtained from Qingdao Makll Group (Qingdao, Shandong, China). Silica gel was first activated by reflux in conc hydrochloric acid for 4 h to remove any adsorbed metal ions, then filtered, washed repeatedly with doubly distilled water to neutral filtrate and dried in an oven at 160°C for 8 h to remove surface water.
Stock standard solutions (1 g L^-1) of each standard were prepared in acetonitrile. Standard solutions were prepared daily by diluting stock solutions in CHCl to concentrations of 0.02, 0.05, 0.1, 0.15, 0.2, 0.3, 0.4 and 0.5 g L^-1 ofloxacin. All solutions were stored in the dark at -4º

Fig. 1 The chemical structures of ofloxacin, enrofloxacin and flumequine

2.2 Instrumentation and chromatographic conditions
A Shimadzu (Kyoto, Japan) LC-6A Liquid Chromatographic System with UV detector with a spectral range from 200 to 600 nm was used to analyze the tested solutions. The chromatographic analysis were performed on a Kromasil C18 column (250 mm×4.6 mm I.D., particle size 5 m m). The mobile phase was 0.025 mM orthophosphoric acid-acetonitrile (70:30) containing 2.5 mM 1-heptanesulfonic acid and the flow rate was 1 mL min^-1 at ambient temperature. Aliquots of 20 m L were injected into the column and the chromatograms were recorded spectrophotometrically at 284 nm.
2.3 Preparation of MIP
A sacrificial silica imprinting concept ^[4] was applied to the preparation of a new molecularly imprinted polymer based on sacrificial mesoporous (Si-MIPs) for selective solid phase extraction of the quinolone residue from edible animal products.
For pre-polymerisation mixtures preparation, 0.361 g (1 mmol) of template (ofloxacin), 0.34 ml (4 mmol) of MAA and 3.77 ml of EGDMA (20 mmol) were dissolved in 14 ml of chloroform in a 25 ml screw-capped glass vial and sonicated. After complete dissolution, the solution was churn up 30 min and then purged with N for 5 min. In order to obtain discrete silica Cpolymer particles and to prevent sticky particle agglomerates. Finally the solution was allowed to stand for 24 h at 40º
The pre-polymerisation mixture was added to 3.000 g silica and shaken vigorously. The mixture was then gently stirred with a spatula until all the mixture penetrated the silica pores and distributed uniformly throughout the surface of silica. Subsequently, 0.0164 g of AIBN was dissolved in 3 ml of chloroform and added to the mixture. Finally, the mixture was sonicated to remove any entrapped air bubbles. A free-flowing material was then obtained. It was flushed with N, sealed and shaked in a mechanical shaker at 60ºfor 24 hour. After polymerisation was completed, Si-MIP composite materials were obtained.
The obtained composite materials was crushed, ground and wet-sieved with acetone. The particle size fraction of 54-74 m m was collected. The resulting particles were transferred into a screw-capped polypropylene tube. The suspension was then cooled in a water-ice bath and 30 ml of 40% aqueous HF was added in several portions while shaking the tube. The suspension was allowed to incubate overnight on a rocking table at room temperature to completely dissolution of the silica matrix of the composite. After dissolution of the silica, the suspension was diluted with 100 ml deionised water, filtered on a glass filter funnel (7 m m pores) and washed extensively with 2 L of deionised water until neutrality. The particles-free HF were placed in a Soxhlet extraction apparatus and washed with ethanol/glacial acetic acid (8:2, v/v ) solution until ofloxacin could no longer be detected at 284 nm in the eluent. Then the particles were washed with methanol to remove residual acetic acid and dried to constant weight under vacuum at 60ºC. A non-molecularly imprinted polymer (NMIP) was prepared in the absence of template and a molecularly imprinted polymer in the absence of Mesoporous Silica Templates (MIP), and treated in an identical manner.
2.4 The scanning electron microscopy
A scanning electron microscopy (SEM) was employed for the analysis of the characteristics of MIP and NMIP. SEM analysis was carried out metal coating at room temperature by using low vaccum mode.
2.5 MISPE cartridges preparation,washing, and elution procedures
A slurry of 100 mg of the Si-MIP or the control polymer (CTL) in 2 mL of MeOH, respectively, was packed into 1 mL SPE syringe barrels and capped with fritted polyethylene disks at the top and at the bottom. Prior to each use, the columns were washed successively with 10%(v/v) acetic acid/methanol (5 mL), acetonitrile (2 ml) and chloroform (2 ml). As a control, a blank SPE column was also prepared in the same manner but with the blank polymer.
A 10 ml sample spiked at 100 mg L^-1 with a mixture of fluoroquinolones OFL, ENR and FUL was prepared in chloroform solvent was passed through at a flow rate of 1 mL min^-1, then the cartridge was washed with 5 ml of acetic acid in ethanol (1:99). The analytes retained in the cartridge were eluted with 5 mL of acetic acid in methanol (10:90). The collected aliquots from the washing step were directly analyzed by HPLC. The collected aliquots from the elution step (1 mL) were dried under nitrogen and redissolved in 1 mL the mobile phase.
2.6 Pretreatment of the biological sample ^{[13, 14]}
Muscle sample were obtained from chicken, and were stored at -10ºC.A sample of liver and muscle (5.00 g wet mass) was accurately weighed. In order to investigate the effect of spiking procedure on extraction efficiency, the fluoroquinolone standard solutions were prepared in acetonitrile. 5 g of tissue was Spiked with the mixture of quinolone at 100m g kg^-1. The ground muscle or liver was added a mixture containing 10% TCA-acetonitrile (8:2) and then the sample was homogenized. After homogenization the tube content was centrifuged at 3500 g for 10 min (4º). The supernatant was filtered, dried in a gentle stream of nitrogen, redissolved in 10 ml chloroform solvent and passed through the MISPE cartridge. The washing, elution, and nanlytical procedures were the same as described above.

3. RESULTS AND DISCUSSION
3.1 Polymer synthesis and recognition mechanism
Silica particles are excellent materials as HPLC stationary phases. They can be dissolved and removed with appropriate reagents and that makes them very suitable both as a support and sacrificial material for the method presented^[4]. The novel method for preparing Si-MIP based on the sacrificial mesoporous silica with high surface area, uniformly shaped particles and uniform pore size was developed. The synthetic MIPs approach involves a pre-polymerisation mixture penetrated the pores of the silica and accumulated onto the silica surface by electrostatic interactions in aprotic solvent (i.e. chloroform), and the interaction of the template molecule with functional monomers is hydrogen bonds in preparing imprinting polymers, followed polymerisation in the presence of an initiator. Discrete single silica-polymer composite materials with the shape of the silica support (Si-MIP composites) were obtained. In a following step, the silica can be dissolved and removed, resulring in materials consisting of molecularly imprinted polymer.
3.2 Affinity of the imprinted polymer
In order to investigate the binding performance of the Si-MIP, saturation experiments and subsequent Scatchard analysis were carried out^[15]. The sized and washed polymer particles (100 mg) were mixed with a 10 mL chloroform solution of ofloxacin of varied concentration from 50 m mol L^-1 to 1.4 mol L^-1. The mixture was oscillated in a constant temperature bath at 25ºfor 24 h. The mixture was transferred into a centrifuge tube and centrifuged at 4000 rmp for 5 min. The concentrations of free compounds in the solutions were determined by an UV-Spectrometer at 284 nm. The absorption quantity ( ) was calculated by subtracting the free concentrations (C_free) from the initial concentrations (). The average data of triplicated independent results were used for the Scatchard analysis. Binding data can be linearly transformed according to the Scatchard equation Q/C = (Q_max-Q)/K where K is an equilibrium dissociation constant and Q_max is an apparent maximum number of binding sites. When Q/ C_free is plotted versus Q, K and Q_max can be estimated from the slope and the intercept, respectively.
As shown in Fig. 2, the Scatchard plot was not linear, suggesting that the binding sites in Si-MIP are heterogeneous in respect to the affinity for ofloxacin. Because there are two distinct sections within the plot which can be regarded as straight lines, it would be reasonable to assume that the binding sites can be classified into two distinct groups with specific binding properties. Under this assumption the respective K values can be calculated to be 178.9 mg L^-1 and 45.54 mg L^-1, and the respective _max 73.31 mg g^-1 and 40.82 mg g^-1 of dry polymer.

Fig. 2 Scatchard analysis of OFL

3.3 Selectivity of the imprinted sorbent
The structurally similar compound ENR was chosen as the competitive species with OFL for the competitive recognition study. As can be seen in Table 1, distribution coefficient (K), selectivity coefficient of the sorbent () and the relative selectivity coefficient () was obtained in these competitive experiments. Distribution coefficient (K) suggested the character of a substance adsorbed by a sorbent, selectivity coefficient of the sorbent () suggested the otherness of two substances adsorbed by one sorbent and relative selectivity coefficient () suggested the otherness of two sorbents. These factors were calculated as the following formula (1) C(3). OFL and ENR had the similar K on the non-imprinted silica sorbent but Si-MIPs showed about four times adsorbed capacity to OFL than to ENR. The (OFL/ENR) value of the Si-MIP sorbent (3.79) was larger than that of NMIP sorbent (1.13), which showed that the Si-MIP sorbent had high selectivity for OFL over the structurally similar compounds ENR. The value was 3.35, which was greater than 1 and showed the Si-MIPs sorbent had higher selectivity than the NMIPs sorbent.

Table 1Competitive loading of OFL and ENR by Si-MIP and NMIP sorbents

Sorbents	Initial solutions (mg L-1	Final solutions (mg L-1
OFL	ENR	OFL	ENR	OFL	ENR
Si-MIP	500	500	345	447	44.93	11.86	3.79	3.35
NMIP	500	500	482	484	3.73	3.3	1.13

= [( – )/] × [volume of solution (ml) / mass of gel (g)] (1)
and represent the initial and final concentrations
= K_dOFL / K_dENR (2) _imprinted / _nonimprinted (3)

Fig. 3 Dynamics adsorption curve of OFL
MIP: molecular imprinting polymer based on ofloxacin template; Si-MIP: MIP based on ofloxacin and silica-gel template.

3.4 Adsorption dynamics
In a typical uptake kinetics test, 100 mg of Si-MIP was added to 10 mL of 100 mg L^-1 ofloxacin chloroform solution in a 25 ml screw-capped glass vial. The mixture was mechanically shaken for 0.5, 1, 1.5, 2.5, 3 h at room temperature. The mixture was transferred into a centrifuge tube and centrifuged at 4000 rmp for 5 min, then the concentrations of free compounds in the solutions were determined by an UV-Spectrometer at 284 nm Dynamics method was the same as static method except oscillated different times and at the constant concentration (100 mg L^-1). The absorption quantity (Q) was caculated by subtracting the free concentration from the initial concentrations.
As shown Fig. 3, the uptake kinetics of ofloxacin. It is clear that the solid extracted process of Si-MIPs is fairly rapid, the 95% uptake of ofloxacin was acieved within 1.5 h. Compared with Si-MIP, the molecularly imprinted polymers in the absence of mesoporous silica templates adsorb ofloxacin slower, attain 90% at least for 2 h.

Fig.4 SEM of the materials. Si-MIP: based on ofloxacin and silica-gel template; MIP: molecular imprinting polymer based on ofloxacin template.

The sacrificial silica imprinting protocol improve adsorption dynamics performance of imprinting materials, which was also elucidated using the scanning electron microscopy and the micrographs of these materials are shown in Fig. 4. The SEM image of Si-MIP with MIP comparison showed that there were many macropores and uniform pore size distribution in the network skeleton of imprinted polymers, which allow the achievement of very high adsorption-desorption efficiencies.
3.5 Optimisation of the MISPE procedure
The optimisation of the whole MISPE procedure (conditioning, loading and elution) conditions of the steps, a standard solution of OFL was applied to the determination of the selectivity of the Si-MIP in SPE and achieve good recoveries for the analyte of interest^{[10, 11]}. The conditoning and the loading steps were optmised first of all. Thus, 10 mL of sample containing 100 mg L^-1 of three fluoroquinolones (OFL, ENR and FLU) was prepared in chloroform and passed through the MISPE.
The cartridge had been conditioned previously with 10% (v/v) acetic acid/methanol (5 ml ), acetonitrile (2 ml) and chloroform (2 ml), and the leakage of MISPE was checked in each of the MISPE experiments. Chloroform was found to be an optimal sample solvent for all the analytes were retained on the Si-MIP mainly by selective interactions. The quinolone molecules have three active sites forming hydrogen bond with Si-MIP for binding to the Si-MIP during the loading step.
A clean-up step (selective washing step) serve to remove as much as possible of sample contaminants while analyte and enhance the selectivity of Si-MIP. This step first optimised when the sample was applied in chloroform. Thus, toluene, water, methanol ethanol and acetonitrile (ACN) were investigated as potential washing solvents for the clean-up step. The best results were obtained using 5 ml of acetic acid in ethanol (1:99). Under these conditions there was an improvement since ENR and FLU were recovered at a level of 70% and 85%, respectively, whereas OFL remained completely bound to the Si-MIP. Ethanol/acetic acid was used as the wash solvent in all subsequent experiments.
Under acidic conditions, OFL would be deprotonated for rapid desorption. Guided by this hypothesis, several acidic solutions were investigated for the elution. The best conditions found to quantitatively elute and recover all the retained analytes was to use 5 ml of acetic acid / methanol (10:90).
3.6 Determination of quinolones in spiked chicken samples
The spiked tissue sample was extracted according to section 2.6. Subsequently,the supernatant was filtered, dried in a gentle stream of nitrogen, redissolved in chloroform or water and passed through the Si-MIP cartridge, MIP cartridge and C18 cartridge, which previously have been conditioned with 3 ml of methanol followed by 3 ml of water, and elution was accomplished using 2 ml of HPLC mobile phase. Si-MIP cartridge and MIP cartridge were performed. The recoveries, reproducility, and LOD of the tissue extracts were calculated and summarized in Table 2.

Table 2 Recoveries(%), precision, and limits of detection (LOD) of OFL, ENR and FLU after MISPE of Chicken sample (spiked 100 m g/kg)

Compounds	Si-MIP carrtridge	MIP carrtridge	18 carrtridge
Recovery(%)	RSD =3, %)	LOD (m g/kg)	Recovery(%)	RSD =3, %)	LOD (m g/kg)	Recovery(%)	RSD =3, %)	LOD (m g/kg)
ofloxacin	92.5	4.2	18	89.2	5.10	22	63.8	3.63	19
enrofloxacin	30.4	4.7	21	31.6	4.04	23	58.56	4.06	12
flumequin	21.7	3.2	42	20.5	4.71	43	72.2	3.36	16

It is clear that C18 cartridges have poorer recoveries for three quinolones compared with Si-MIP cartridges and MIP cartridges. The mechanism of C18 bonded-phase extraction is based on non-polar interactions between the carbob Chydrogen bonds of the sorbent and the carbon Chydrogen of bonds of the analyte^[13-15]. This confirmed the reliability and efficacy of the proposed method for the analysis ofquinolone residues in real samples.

4. CONCLUSION
In this study, we tested a new technique to synthesize a ofloxacin molecularly imprinted polymer base on silica-gel sacrificial material. As the templates, ofloxacin was used here and binding to the molecularly imprinted polymer was confirmed to be highly selective. This ofloxacin recognition depend on shape of the pores created on polymer.
The approach, therefor, may have great potential application for generation of separate materials for chromatogram and solid phase extraction stationary phase. Using the MIPs as selective solid phase extraction stationary phase, we analyzed the chicken had not found the ofloxacin and enrofloxacin in that samples. The study results presented here have substantiated the significant research interest in molecularly imprinted polymer base on silica-gel sacrificial material due to their ease of preparation, high separation efficiency, and rapid mass transport.

Acknowledgement  This research was supported by National Science Foundation of China (No. 20675024), Natural Science Foundation of Hebei Educational Committee (No. 2006407), China Postdoctoral Science Foundation (No. 2005037629) and Science Foundation of Hebei University.