- Email: [email protected]
Novel dummy molecularly imprinted polymers for matrix solid-phase dispersion extraction of eight fluoroquinolones from fish samples
Journal of Chromatography A, 1359 (2014) 1–7
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Novel dummy molecularly imprinted polymers for matrix solid-phase dispersion extraction of eight fluoroquinolones from fish samples Xiaoli Sun a,b , Jincheng Wang a , Yun Li a , Jiajia Yang a,b , Jing Jin a , Syed Mazhar Shah a , Jiping Chen a,∗ a b
Key Laboratory of Separation Science for Analytical Chemistry, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China University of Chinese Academy of Sciences, Beijing 100049, China
a r t i c l e
i n f o
Article history: Received 5 June 2014 Received in revised form 5 July 2014 Accepted 7 July 2014 Available online 12 July 2014 Keywords: Fluoroquinolones Molecularly imprinted polymers Dummy template Class-selective Matrix solid-phase dispersion.
a b s t r a c t A series of novel dummy molecularly imprinted polymers (DMIPs) were prepared as highly class-selective sorbents for fluoroquinolones. A non-poisonous dummy template, daidzein, was used for the first time to create specific molecular recognition sites for fluoroquinolones in the synthesized polymers. The influence of porogen polarity on dummy molecular imprinting effect was studied. The DMIP prepared using dimethylsulfoxide−acetonitrile (1:1.8, v/v) as porogen achieved the highest imprinting factors (IF) for fluoroquinolones over a range of IF 13.4–84.0. This DMIP was then used for selective extraction of eight fluoroquinolones (fleroxacin, ofloxacin, norfloxacin, pefloxacin, ciprofloxacin, lomefloxacin, enrofloxacin and gatifloxacin) from fish samples based on dummy molecularly imprinted matrix solid-phase dispersion (DMI-MSPD). The extracted fluoroquinolones were subsequently analyzed by high-performance liquid chromatography (HPLC) equipped with a fluorescence detector (FLD). The developed method had acceptable recoveries (64.4−102.7%) and precision (RSDs: 1.7−8.5%, n = 5) for determination of fluoroquinolones in fish samples fortified at levels of 10 and 100 ng g−1 . The limits of detection (LODs) for identification of eight fluoroquinolones ranged between 0.06 and 0.22 ng g−1 . The results demonstrated great potential of the optimized method for sample preparation in routine analysis of trace fluoroquinolones in fish samples. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Fluoroquinolones (FQs) are one of the most important classes of synthetic antibiotics, which show excellent activity against pathogenic Gram-negative and Gram-positive bacteria, as well as anaerobes [1]. The extensive use of fluoroquinolones in foodproducing animals has raised concerns about the potential risks of their residues in foodstuffs of animal origin [2] or environmental matrix [3–5]. These residues can be directly toxic or can elevate the pathogen resistance levels and possible allergic hypersensitivity reactions in humans [6,7]. Maximum residue limits (MRLs) for several FQs have been established in many countries for animal foods [8]. Therefore, monitoring of FQ residues in livestock, poultry, fish and other animal products for human consumption is greatly significant. Currently, the most used methods for fluoroquinolone analysis are based on high-performance liquid chromatography (HPLC),
∗ Corresponding author. Tel.: +86 411 84379562; fax: +86 411 84379562. E-mail address: [email protected] (J. Chen). http://dx.doi.org/10.1016/j.chroma.2014.07.007 0021-9673/© 2014 Elsevier B.V. All rights reserved.
mainly coupled with fluorescence, ultraviolet (UV) or mass spectrometric (MS) detection [9–11]. Different sample preparation methods have been developed for the analysis of fluoroquinolones in biological samples, such as liquid extraction (LE), protein precipitation (PP), liquid–liquid extraction (LLE), solid-phase extraction (SPE), supercritical fluid extraction (SFE) and so on [8,12]. The main limitations of these methods include the relatively low recoveries, the tedious and time-consuming extraction and cleanup processes, the use of a large amount of toxic organic solvents and the low specificity toward fluoroquinolones. Hence alternative sample preparation methods which can tackle these issues are highly desirable for the extraction and enrichment of fluoroquinolones. Matrix solid-phase dispersion (MSPD), first introduced in 1989 by Barker et al. [13], is one of the most promising techniques for the simultaneous disruption, extraction and cleanup of solid, semisolid or viscous samples [14]. MSPD involves blending a viscous, solid or semi-solid sample with an appropriate sorbent (silica, florisil, alumina, C18, C8, etc.) followed by packing, washing and elution. In MSPD, the procedures of homogenization, disruption, extraction and cleanup are combined into one simple process, thus
2
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
greatly reducing analysis time and solvent use and increasing sample throughput. In recent years, MSPD has been widely used in the pre-treatment of food, plant, animal tissues, human biological samples, environmental samples and cosmetics [15]. However, due to the lack of special selectivity, MSPD using traditional sorbents was confronted with the difficulty of selectively extracting target analytes from complex samples. Molecular imprinting is a rapidly developing technique for preparation of polymers with excellent molecular recognition properties, which has been found effective in a variety of extraction techniques such as molecular imprinted solid-phase extraction (MI-SPE), molecular imprinted solid-phase microextraction (MISPME), molecular imprinted stir-bar sorptive extraction (MI-SBSE), molecular imprinted matrix solid-phase dispersion (MI-MSPD) [16] and so on. Fluoroquinolone-imprinted polymers have been applied for their enrichment and purification from complex samples such as milk [17], eggs [18], urine [19], fish [20], serum [21], water [22,23] and soil [24]. Good recoveries as well as high selectivity have been obtained due to the good molecular recognition properties of molecularly imprinted polymers (MIPs). However, all the FQs-MIPs reported previously were prepared using one (pefloxacin [25], gatifloxacin [26], ciprofloxacin, ofloxacin [27] and enrofloxacin) or two (levofloxacin−ciprofloxacin [28]) of fluoroquinolones as templates, where possible leakage of template molecules still happened even after exhaustive washing steps. Template leakage could have a serious impact on the accuracy of analytical method [29] or made it unsuitable for simultaneous analysis of the whole class of fluoroquinolones. This problem has become one of the major areas of concern in sample pretreatment methods employing MIPs [30]. The use of a dummy molecule presents an easy solution to circumvent this problem since any leakage will be different from the analyte [31]. Until now, there has been no report about the use of dummy template that does not belong to fluoroquinolones for the preparation of dummy molecularly imprinted polymers (DMIPs) for fluoroquinolones. This work presents the first attempt of using daidzein as a non-poisonous dummy template for the imprinting of fluoroquinolones. The synthesized DMIPs were used as selective MSPD sorbents for simultaneous determination of eight fluoroquinolones in fish samples. The selectivity, recovery and precision of the developed method were also evaluated.
2. Experimental 2.1. Chemicals and reagents Daidzein was obtained from Zhongxin Pharmaceuticals (Tianjin, China). Enoxacin (ENO), norfloxacin (NOR), ciprofloxacin (CIP), pefloxacin (PEFX), ofloxacin (OFL), lomefloxacin (LOM), fleroxacin (FLX), enrofloxacin (ENR) and gatifloxacin (GAT) were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). The structures of these fluoroquinolones and daidzein are shown in Fig. 1. Diethylstilbestrol (DES), ethylene dimethacrylate (EDMA) and trifluoroacetic acid (TFA) were purchased from J&K Chemical Ltd. (Beijing, China). The initiator 2,2 -azobisisobutyronitrile (AIBN) was supplied by Aladdin Chemical (Shanghai, China). 4-Vinylpyridine (4-VP) was obtained from Acros (NJ, USA). HPLC-grade acetonitrile and methanol were purchased from Fisher (Fair Lawn, NJ, USA). Dimethylsulfoxide (DMSO) and tetrahydrofuran (THF) were purchased from Tianjin Bodi Chemical Engineering Co., Ltd. (Tianjin, China). Dimethylformamide (DMF) was obtained from Tianjin Damao Chemical Reagent Factory (Tianjin, China). Purified water by Milli-Q Plus water purification system (Millipore, Bedford, with a 70 mm × 0.22 m filter) was used throughout the experiments.
2.2. Preparation of imprinted and non-imprinted polymers The dummy molecularly imprinted polymers were prepared using different porogens. The composition of the prepolymerization mixtures are presented in Table 1. The dummy template daidzein (0.2542 g, 1 mmol) was dissolved in 5.6 mL of porogen in a 10 mL thick-walled glass tube. The functional monomer 4-VP (0.42 mL, 4 mmol), cross-linking monomer (EGDMA) (3.8 mL, 20 mmol) and the initiator (AIBN) (0.04 g) were then successively added to the above solution. The solution was sonicated and saturated with dry nitrogen for 10 min before the glass tube was sealed. The tube was then placed in a water bath at 60 ◦ C for 24 h. The obtained DMIPs were crushed, ground, sieved (particle size range: 38.5–63.0 m), sedimented in acetone and dried under vacuum. The template was extracted by extensive washing with methanol−acetic acid (9:1, v/v). The non-imprinted
Fig. 1. Chemical structures of the dummy template, structure analogue and selected fluoroquinolones.
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7 Table 1 Composition of the polymerization mixtures used for the preparation of the dummy molecularly imprinted polymers (DMIPs) and corresponding non-imprinted polymers (NIPs). Polymer DMIP1 NIP1 DMIP2 NIP2 DMIP3 NIP3 DMIP4 NIP4
Daidzein (mmol)
4-VP (mmol)
EGDMA (mmol)
Porogen
1 — 1 — 1 — 1 —
4 4 4 4 4 4 4 4
20 20 20 20 20 20 20 20
DMSO DMSO DMSO:ACN (1:1.8, v/v) DMSO:ACN (1:1.8, v/v) DMF:ACN (1:1.8, v/v) DMF:ACN (1:1.8, v/v) DMF:THF (1:1.8, v/v) DMF:THF (1:1.8, v/v)
polymers (NIP) were prepared simultaneously using the same protocol in the absence of the template molecule. 2.3. Chromatographic evaluation of the polymers The DMIPs and NIPs were suspended in methanol, sonicated in a water bath and then slurry packed into stainless steel HPLC columns (100 × 4.6 mm i.d.) at 3000 psi using an air-driven fluid pump (Haskel, Burbank, CA, USA) with ethanol as the pushing solvent. The chromatographic evaluation of the polymers was carried out using a Waters 515 HPLC pump equipped with a Waters 2487 dual absorbance detector. The binding affinities of polymers toward dummy template were evaluated with acetonitrile as mobile phase and daidzein as analyte. Then, the cross-selectivity of the polymers toward fluoroquinolones was investigated using methanol as mobile phases and enoxacin, fleroxacin, ofloxacin, norfloxacin, pefloxacin, ciprofloxacin, lomefloxacin, enrofloxacin, gatifloxacin, diethylstilbestrol and daidzein as analytes. The results were presented and discussed in Sections 3.1.1 and 3.1.2, respectively. A 20 L aliquot of the analyte (20 ppm) was injected for the analysis with a flow rate of 1 mL min−1 . The UV detector was set at 280 nm. Acetone was injected as a void marker. The capacity factor (k) was calculated as k = (tR − t0 )/t0 , where tR and t0 are the retention times of the analyte and the void marker, respectively. The molecular imprinting factor (IF) was calculated as IF = kMIP /kNIP , where kMIP and kNIP are the capacity factors of the analyte on the imprinted and non-imprinted polymers, respectively. 2.4. MSPD procedure The fish samples (about 500 g) used for this study were collected from a local market. The fish muscle (edible parts) was minced with a meat grinder and stored in well-sealed containers at −20 ◦ C before analysis. One hundred and fifty milligrams of DMIP2 was placed in a glass mortar, and 200 mg of fish meat was added onto the DMIP2. An appropriate volume of fluoroquinolone standards was injected randomly into the fish meat and allowed to stand for 60 min at room temperature. Blank samples were prepared similarly, except that the same volume of methanol containing no fluoroquinolone was added. After adding 200 L of water, the fish meat was blended with the DMIP2 material using a glass pestle until a homogeneous mixture was achieved (∼5 min). The homogenized sample was introduced into an SPE cartridge packed with 50 mg of DMIP2 (equilibrated with 2.0 mL of acetonitrile and 2.0 mL of water). The column was then tapped slightly to remove any air pocket. After rinsing with 3.0 mL of methanol−water (20:80, v/v), the fluoroquinolones were eluted with 4.0 mL of acetonitrile−trifluoroacetic acid (99:1, v/v). The eluents were dried under a moderate stream of nitrogen gas and then reconstituted in 500 L of methanol−water (50:50, v/v). An aliquot of 20 L was injected into the HPLC system for analysis.
3
Table 2 The dielectric constant (ε), Snyder polarity index (SPI), the capacity factors (k) and imprinting factors (IF) of DMIP1, DMIP2, DMIP3 and DMIP4 towards dummy template (daidzein). Polymer
Porogen
εa
SPIb
kMIP
kNIP
IF
DMIP1 DMIP2 DMIP3 DMIP4
DMSO DMSO:ACN (1:1.8, v/v) DMF:ACN (1:1.8, v/v) DMF:THF (1:1.8, v/v)
46.70 39.37 36.89 18.90
7.20 6.20 5.96 4.89
9.99 16.56 27.80 33.23
5.22 4.11 5.19 5.22
1.9 4.0 5.4 6.4
ε and SPI values of mixed solvents were calculated by the equation: εmix =
ϕi εi
and SPImix = ϕi SPIi . Coefficients ϕi represent the mole fraction of solvent. The ε and SPI values of pure DMSO, DMF, ACN and THF were getting from the literatures [32–34]. a Measured dielectric constant. b Snyder polarity index [35].
2.5. HPLC analysis HPLC analysis was performed on an Agilent 1200 system (Agilent Technologies, Palo Alto, CA, USA) equipped with a vacuum degasser, a quaternary pump, an auto-sampler and a fluorescence detector (FLD) connected to a reversed-phase column (Agilent ZORBAX SB-C18, 250 × 4.6 mm i.d., particle size, 5 m). A gradient program was used at a flow rate of 1 mL min−1 , by combining solvent A of water−formic acid (99.9:0.1, v/v), solvent B of methanol and solvent C of acetonitrile as follows: 9.5−11.5% C (24 min), 11.5−26.5% C (10 min), 26.5−96.5% C (2 min), 96.5% C (5 min). Solvent B was maintained at 3.5% during the whole separation process. The column temperature was kept at 15 ◦ C. The injection volume was 20 L, and the FLD excitation and emission wavelengths were set at 290 and 480 nm, respectively. 3. Results and discussion 3.1. Synthesis and chromatographic evaluation of the DMIPs The molecularly imprinted polymers for fluoroquinolones (FQs) were prepared by using daidzein as the dummy template to avoid the template bleeding problem. Daidzein can be dissolved well only in organic solvents such as DMSO and DMF. In order to improve the imprinting effect, less polar solvents such as ACN and THF were added into DMSO or DMF up to a maximum proportion while still maintaining good solubility for the dummy template. The binary mixed solvents with different polarities were used as porogens for the polymerization and their effects on the affinity and selevtivity of the synthesised DMIPs toward dadzein and FQs were investigated. 3.1.1. Effect of porogen polarity on the selectivity of DMIPs toward daidzein The DMIPs were first evaluated in terms of their capacity factors (kMIP ) and imprinting factors (IF) toward the dummy template molecule, daidzein. ACN was used as the mobile phase. The polarity of porogen solvent is described by both Snyder polarity index (SPI) and dielectric constant (ε). The relationship between porogen polarities and the imprinting factors of daidzein on DMIPs (DMIP1, DMIP2, DMIP3 and DMIP4) was discussed. As can be seen from Table 2, DMIP4 synthesized from the porogen with the lowest SPI and ε values achieved the highest capacity and imprinting factor for daidzein. Negative correlations between the SPI/ε values and the IF values of DMIPs were obtained by further correlation analysis and linear fitting, as shown in Fig. 2. SPI was found to be a more adequate parameter to describe the porogen polarity with a square value of correlation coefficient (r2 ) of 0.93. These results can be associated with the stronger interactions between basic -donor/acceptor monomer (4-VP) and daidzein in
4
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
Fig. 2. Influence of dielectric constant (ε) and Snyder polarity index (SPI) of the porogens on the imprinting factors (IF) of DMIPs towards daidzein.
less polar solvents, thus creating more high-affinity binding sites during the polymerization [36]. 3.1.2. Effect of porogen polarity on the cross-selectivity of DMIPs toward fluoroquinolones The dummy molecular imprinting effects of DMIPs toward fluoroquinolones were investigated by using acetonitrile, methanol, water or their mixture as the mobile phases. The best selectivity for fluoroquinolones was found when methanol was used, and the resulted kNIP , kMIP and IF values are listed in Fig. 3. As can be seen in Fig. 3A–D, DMIP1 and DMIP2 showed higher capacities and imprinting factors for fluoroquinolones as compared to
DMIP3 and DMIP4. This observation was significantly different from the case for daidzein. The negative correlations between SPI/ε and the imprinting factors for daidzein no longer exist for fluoroquinolones. The most plausible reason of these results is the different cavity sizes produced from daidzein template by using porogens with different polarities. Porogens with lower polarities can lead to the formation of more rigid cavities for daidzein, which are not easily accessible to larger fluoroquinolone molecules. Therefore, the strong acid–base electrostatic interactions between 4-VP and fluoroquinolones in the cavities are disabled, resulting in very low affinities. DMIP2 with suitable imprinted cavity size by using dimethylsulfoxide−acetonitrile (1:1.8, v/v) as the porogen achieved the highest imprinting factors for fluoroquinolones over a range of 13.4–84.0. Furthermore, we can see from Fig. 3B that the capacity factors of different fluoroquinolones on DMIP2 were inversely related to their molecular sizes. The smallest ENO and NOR molecules among the fluoroquinolones obtained the highest capacity factors. The methyl groups of OFL, LOM and PEFX on piperazine rings situated at position 7 reduced their capacity factors to some degree due to their bulky molecules. In addition, the large ethyl fluoride or cyclopropyl ring groups of FLX, CIP, ENR and GAT dramatically reduced their retention abilities on the DMIP2. In summary, porogen polarity plays a very important role in the affinity and selectivity of dummy molecular imprinting. Highly efficient imprinting of dummy template does not always result in high affinities of target molecules when the structures of dummy template and target molecules are not exactly the same. The recognition abilities of DMIP2 for fluoroquinolones presented here were excellent and the IF values tested were much higher than those reported in literatures [23,26]. Finally, low affinities were observed
Fig. 3. Evaluation of the cross-selectivity of DMIPs in terms of capacity factors (k) and imprinting factors (IF) towards ENO, FLX, OFL, NOR, PEFX, CIP, LOM, ENR, GAT, DAI and DES: (A) DMIP1, (B) DMIP2, (C) DMIP3, (D) DMIP4. Conditions: column, 150 mm × 4.6 mm; mobile phase, methanol; flow rate, 1 mL min−1 ; sample volume, 20 L (20 ppm); UV detection, 280 nm (see Section 2.3). Acetone was used as the void marker.
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
5
Fig. 4. MSPD extraction recoveries (%) obtained with the DMIP2 and the NIP sorbents for eight fluoroquinolones from fish samples spiked at 100 ng g−1 level using a washing step with 3 mL of methanol−water (20:80, v/v).
for daidzein and DES when methanol was used as the mobile phase with imprinting factors of 1.08 and 1.29, respectively, which further confirmed the good class selectivity of DMIP2 for fluoroquinolones. 3.2. Optimization of DMI-MSPD procedure 3.2.1. Effect of the mass ratio of sample to DMIP2 The first step for the development of an MSPD method was selecting a suitable mass ratio of sample/sorbent, in order to allow complete adsorption of the sample components and to facilitate the sample transfer onto the cartridge. The sample/sorbent ratios typically ranged from 1:1 to 1:4 using traditional SPE sorbents in the previously reported works [37]. Another research reported the optimal sample/sorbent ratio of 1:1 using FQs-MIPs as sorbents [18]. In our study, different sample/sorbent ratios were evaluated (2:1, 4:3, 1:1 and 2:3) using 200 mg of the sample. The results showed that a sample/sorbent ratio of 4:3 was sufficient for complete retention of fluoroquinolones, with recoveries ranging from 66.8 to 104.2% for triplicate experiments. Further increasing the proportion of sorbents gave no improvement for the recoveries of fluoroquinolones. However, less DMIP2 sorbents resulted in heterogeneous sample mixture and low recoveries (e.g. 48.2−86.7% for 2:1). Therefore, the sample/sorbent ratio of 4:3 was finally selected for further investigations. 3.2.2. Effect of washing and eluting conditions The nature and volume of washing and elution solvent are the key factors, which should be carefully optimized to achieve high recoveries for analytes, while eliminating most of the interferences originating from a complex biological matrix. The effect of washing solvent type and volume on recoveries and selectivity was studied based on 150 mg of DMIP2 sorbent and 200 mg of fish sample (100 ng g−1 of spiking level for each fluoroquinolone). According to Section 3.1.2, high selectivity of DMIP2 for fluoroquinolones can be achieved by using methanol as the mobile phase. Therefore, methanol was used as the selective washing solvent at first. Unfortunately, despite its high selectivity, the use of methanol led to low recoveries due to its strong elution strength. When 2 mL of methanol was used, the recoveries less than 40% were obtained on DMIP2. The mixtures of water−methanol (methanol content ranging from 0 to 50%) were tested. High methanol content resulted in low recoveries of fluoroquinolones. When 2 mL of water−methanol (50:50, v/v) was used, the recoveries from 36.4 to 84.8% were obtained. In order to achieve higher recoveries, water−methanol (80:20, v/v) was finally selected. Different volumes of water−methanol (80:20, v/v) were investigated from 2.0 to 4.0 mL, and 3.0 mL was found to be the optimum volume of washing solvent. The results are shown in Fig. 4.
Fig. 5. Chromatograms of spiked fish samples (10 ng g−1 ). Blue line directly extracted fish sample (200 mg of fish sample extracted by 4 mL of acetonitrile−TFA (99:1, v/v) in an ultrasonic bath for 5 min); red line fish sample after DMI-MSPD process. Peaks: 1, FLX; 2, OFL; 3, NOR; 4, PEFX; 5, CIP; 6, LOM; 7, ENR; 8, GAT.
In this case, the extraction recoveries for eight fluoroquinolones on the NIP ranged between 32.4 and 86.1%, whereas those on DMIP2 ranged between 66.8 and 104.2%. These results further confirmed the existence of the specific interactions between fluoroquinolones and DMIP2. The low selectivity of DMIP2 for ENR and PEFX was mostly due to the high hydrophobic interactions (nonspecific) resulted from their relatively high octanol/water partition coefficients (log Kow ). The log Kow values are 0.22 and 0.31 for ENR and PEFX, respectively, while those of the other fluoroquinolones are all below zero [38]. The final elution of fluoroquinolones was conducted by using methanol−ammonia (99:1, v/v), methanol−TFA (99:1, v/v) and acetonitrile−TFA (99:1, v/v). The best recoveries were obtained by using acetonitrile−TFA (99:1, v/v) as elution solvent with a relatively clean baseline. 3.3. Method validation Two calibration curves were established for each fluoroquinolone to cover a wide concentration range 2.5−400 ng g−1 , i.e. a low curve for concentrations below 30 ng g−1 and a high curve for concentrations above 30 ng g−1 . Good linearity was obtained for all analytes in both low and high concentrations, with regression coefficients R > 0.999 (Table 3). To evaluate the applicability of the optimized DMI-MSPD-HPLCFLD procedure to real samples, accuracy and precision of the method were evaluated using spiked blank fish samples at two concentration levels (10 and 100 ng g−1 ). No background of FQs was found in the blank fish samples. The results summarized in Table 3 show acceptable recoveries, repeatability and intermediate precision of the method. For repeatability, five extractions were performed on the same day under the optimum conditions, and the RSD values ranged from 1.7 to 5.7%. The intermediate precision was determined by measuring the RSDs of experiments carried out in different days (n = 5), and day-to-day precision was between 2.4 and 8.5%. The average recoveries of fluoroquinolones at two different levels were ranged from 64.4 to 102.7%. Moreover, as shown in Fig. 5, the use of DMIP2 resulted in less matrix interferences compared to the directly extracted sample, and none of co-elution peaks was observed in the chromatogram. The recoveries obtained for all the fluoroquinolones except LOM were excellent and better than those in literatures using other MIP materials or traditional SPE sorbents [19,20,39]. The limits of detection (LODs) of the DMIMSPD method, calculated as a signal-to-noise ratio of 3, were found
6
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
Table 3 Correlation coefficients, repeatability, intermediate precision, limits of detection and recovery study of the proposed method. Analyte
FXL OFL NOR PEFX CIP LOM ENR GAT a b
Correlation coefficient R2
Repeatability RSD (%, n = 5)
Low curvea
High curveb
10 ng g−1
100 ng g−1
10 ng g−1
100 ng g−1
0.9999 0.9997 0.9991 0.9996 0.9999 0.9999 0.9999 0.9999
0.9996 0.9994 0.9998 0.9993 0.9999 0.9996 0.9993 0.9998
2.0 5.0 3.8 4.1 5.7 2.0 3.9 4.8
1.7 3.2 5.3 2.7 2.7 5.0 2.5 2.6
3.5 7.7 2.6 2.6 2.4 7.9 8.5 6.3
5.5 6.5 5.1 5.7 7.0 5.2 6.5 3.9
Intermediate precision RSD (%, n = 5)
LOD S/N = 3 (ng g−1 )
0.16 0.20 0.14 0.22 0.21 0.06 0.06 0.07
Average recoveries (%) 10 ng g−1
100 ng g−1
93.1 96.7 99.2 94.3 102.7 88.4 99.1 100.0
89.8 98.0 95.0 99.4 96.4 64.4 95.2 90.2
The calibration curve for low concentration levels (2.5, 5, 10, 20, 30 ng g−1 ). The calibration curve for high concentration levels (30, 50, 100, 200, 400 ng g−1 ).
to be in the range 0.06−0.22 ng g−1 (Table 3). The values obtained are lower than those mentioned in previous papers [19,20,39] and are well below the maximum residue levels (MRLs) set by the European Union [8]. The good recoveries, low LODs and excellent accuracy demonstrated the applicability of the DMIP sorbent for pre-treatment of fluoroquinolones in fish samples. 4. Conclusions In this work, highly class-selective dummy molecularly imprinted polymers (DMIPs) for fluoroquinolones were successfully synthesized by using a non-poisonous dummy template molecule (daidzein). The porogen polarity has a significant influence on the dummy molecular imprinting effect. Among the polymers prepared, the DMIP prepared using dimethylsulfoxide−acetonitrile (1:1.8, v/v) as porogen showed the highest imprinting factors for eight fluoroquinolones. A simple and highly selective sample preparation method based on the DMIP for simultaneous determination of eight fluoroquinolones in fish samples was developed by DMIP-based MSPD followed by HPLC-FLD. The new DMI-MSPD-HPLC-FLD method was not only highly selective and accurate, but it also effectively eliminated the template bleeding problem. These features demonstrate great potential of the optimized method for selective extraction of fluoroquinolones from fish samples at the ng g−1 level. Acknowledgments This work was supported by the National Natural Science Foundation of China (Nos. 21277139, 21205117) and National High Technology Research and Development Program of China (863 program, No. 2013AA065203). References [1] S.A. Brown, Fluoroquinolones in animal health, J Vet Pharmacol Ther 19 (1996) 1–14. [2] M.D. Barton, Antibiotic use in animal feed and its impact on human health, Nutr Res Rev 13 (2000) 279–300. [3] Y. Pico, V. Andreu, Fluoroquinolones in soil—risks and challenges, Anal Bioanal Chem 387 (2007) 1287–1299. [4] H. Sanderson, D.J. Johnson, T. Reitsma, R.A. Brain, C.J. Wilson, K.R. Solomon, Ranking and prioritization of environmental risks of pharmaceuticals in surface waters, Regul Toxicol Pharm 39 (2004) 158–183. [5] P. Sukul, M. Spiteller, Fluoroquinolone antibiotics in the environment, Rev Environ Contam Toxicol 191 (2007) 131–162. [6] J. Davies, D. Davies, Origins and evolution of antibiotic resistance, Microbiol Mol Biol Rev 74 (2010) 417–433. [7] A.D. Anderson, J.M. Nelson, S. Rossiter, F.J. Angulo, Public health consequences of use of antimicrobial agents in food animals in the United States, Microb Drug Resist 9 (2003) 373–379. [8] V. Andreu, C. Blasco, Y. Picó, Analytical strategies to determine quinolone residues in food and the environment, Trends Anal Chem 26 (2007) 534–556.
[9] M. Farré, D. Barceló, Analysis of emerging contaminants in food, Trends Anal Chem 43 (2013) 240–253. [10] M. Seifrtova, L. Novakova, C. Lino, A. Pena, P. Solich, An overview of analytical methodologies for the determination of antibiotics in environmental waters, Anal Chim Acta 649 (2009) 158–179. [11] J. Sousa, G. Alves, J. Abrantes, A. Fortuna, A. Falcao, Analytical methods for determination of new fluoroquinolones in biological matrices and pharmaceutical formulations by liquid chromatography: a review, Anal Bioanal Chem 403 (2012) 93–129. [12] J.Y. Shen, M.R. Kim, C.J. Lee, I.S. Kim, K.B. Lee, J.H. Shim, Supercritical fluid extraction of the fluoroquinolones norfloxacin and ofloxacin from orally treated-chicken breast muscles, Anal Chim Acta 513 (2004) 451–455. [13] S.A. Barker, A.R. Long, C.R. Short, Isolation of drug residues from tissues by solid phase dispersion, J Chromatogr A 475 (1989) 353–361. [14] S.A. Barker, Matrix solid-phase dispersion, J Chromatogr A 885 (2000) 115–127. [15] A.L. Capriotti, C. Cavaliere, A. Laganà, S. Piovesana, R. Samperi, Recent trends in matrix solid-phase dispersion, Trends Anal Chem 43 (2013) 53–66. [16] E. Turiel, A. Martin-Esteban, Molecularly imprinted polymers for sample preparation: a review, Anal Chim Acta 668 (2010) 87–99. [17] Y. Lv, L. Yang, X. Liu, Z. Guo, H. Sun, Preparation and evaluation of a novel molecularly imprinted hybrid composite monolithic column for online solidphase extraction coupled with HPLC to detect trace fluoroquinolone residues in milk, Anal Methods 5 (2013) 1848–1855. [18] H. Yan, F. Qiao, K.H. Row, Molecularly imprinted-matrix solid-phase dispersion for selective extraction of five fluoroquinolones in eggs and tissue, Anal Chem 79 (2007) 8242–8248. [19] E. Benito-Pena, S. Martins, G. Orellana, M.C. Moreno-Bondi, Water-compatible molecularly imprinted polymer for the selective recognition of fluoroquinolone antibiotics in biological samples, Anal Bioanal Chem 393 (2009) 235–245. [20] W. Junping, P. Mingfei, F. Guozhen, W. Shuo, Preparation of a novel molecularly imprinted polymer by a sol-gel process for on-line solid-phase extraction coupled with high performance liquid chromatography to detect trace enrofloxacin in fish and chicken samples, Microchim Acta 166 (2009) 295–302. [21] D. Xiao, P. Dramou, N. Xiong, H. He, H. Li, D. Yuan, H. Dai, Development of novel molecularly imprinted magnetic solid-phase extraction materials based on magnetic carbon nanotubes and their application for the determination of gatifloxacin in serum samples coupled with high performance liquid chromatography, J Chromatogr A 1274 (2013) 44–53. [22] A. Prieto, S. Schrader, C. Bauer, M. Moeder, Synthesis of a molecularly imprinted polymer and its application for microextraction by packed sorbent for the determination of fluoroquinolone related compounds in water, Anal Chim Acta 685 (2011) 146–152. [23] E. Benito-Pena, J.L. Urraca, B. Sellergren, M. Cruz Moreno-Bondi, Solid-phase extraction of fluoroquinolones from aqueous samples using a watercompatible stochiometrically imprinted polymer, J Chromatogr A 1208 (2008) 62–70. [24] E. Turiel, A. Martin-Esteban, J.L. Tadeo, Molecular imprinting-based separation methods for selective analysis of fluoroquinolones in soils, J Chromatogr A 1172 (2007) 97–104. [25] M. Zheng, R. Gong, X. Zhao, Y. Feng, Selective sample pretreatment by molecularly imprinted polymer monolith for the analysis of fluoroquinolones from milk samples, J Chromatogr A 1217 (2010) 2075–2081. [26] D. Xiao, P. Dramou, N. Xiong, H. He, D. Yuan, H. Dai, H. Li, X. He, J. Peng, N. Li, Preparation of molecularly imprinted polymers on the surface of magnetic carbon nanotubes with a pseudo template for rapid simultaneous extraction of four fluoroquinolones in egg samples, Analyst 138 (2013) 3287–3296. [27] F. Qiao, H. Sun, Simultaneous extraction of enrofloxacin and ciprofloxacin from chicken tissue by molecularly imprinted matrix solid-phase dispersion, J Pharm Biomed Anal 53 (2010) 795–798. [28] P. Liu, J. Shen, L. Gao, L. Liu, R. Li, Q. Li, Determination of fluoroquinolones in milk by high-performance liquid chromatography using mixed-templates imprinted polymer extraction, Asian J Chem 22 (2010) 6275–6288.
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7 [29] F.G. Tamayo, E. Turiel, A. Martin-Esteban, Molecularly imprinted polymers for solid-phase extraction and solid-phase microextraction: recent developments and future trends, J Chromatogr A 1152 (2007) 32–40. [30] L. Chen, S. Xu, J. Li, Recent advances in molecular imprinting technology: current status, challenges and highlighted applications, Chem Soc Rev 40 (2011) 2922–2942. [31] L. Andersson, A. Paprica, T. Arvidsson, A highly selective solid phase extraction sorbent for pre-concentration of sameridine made by molecular imprinting, Chromatographia 46 (1997) 57–62. [32] P.B. Wright, A.S. Lister, J.G. Dorsey, Behavior and use of nonaqueous media without supporting electrolyte in capillary electrophoresis and capillary electrochromatography, Anal Chem 69 (1977) 3251–3259. [33] T. Liu, I.A. Nicholls, T. Oberg, Comparison of theoretical and experimental models for characterizing solvent properties using reversed phase liquid chromatography, Anal Chim Acta 702 (2011) 37–44.
7
[34] X. Li, R. Chen, M. Huang, M. Zhu, Q. Chen, Synthesis of a soluble pyrrole copolymer with phenetidine, J Polym Sci A Polym Chem 42 (2004) 2073–2092. [35] L. Snyder, Classification off the solvent properties of common liquids, J Chromatogr Sci 16 (1978) 223–234. [36] S. Pardeshi, R. Dhodapkar, A. Kumar, Influence of porogens on the specific recognition of molecularly imprinted poly(acrylamide-co-ethylene glycol dimethacrylate), Compos Interfaces 21 (2014) 13–30. [37] E.M. Kristenson, U.A.T. Brinkman, L. Ramos, Recent advances in matrix solidphase dispersion, Trends Anal Chem 25 (2006) 96–111. [38] D.L. Ross, C.M. Riley, Aqueous solubilities of some variously substituted quinolone antimicrobials, Int J Pharm 63 (1990) 237–250. ˜ ˜ Simultaneous [39] A. Espinosa-Mansilla, A. Jiménez Girón, A. Munoz de la Pena, determination of the residues of fourteen quinolones and fluoroquinolones in fish samples using liquid chromatography with photometric and fluorescence detection, Czech J Food Sci 30 (2012) 74–82.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Novel dummy molecularly imprinted polymers for matrix solid-phase dispersion extraction of eight fluoroquinolones from fish samples Xiaoli Sun a,b , Jincheng Wang a , Yun Li a , Jiajia Yang a,b , Jing Jin a , Syed Mazhar Shah a , Jiping Chen a,∗ a b
Key Laboratory of Separation Science for Analytical Chemistry, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China University of Chinese Academy of Sciences, Beijing 100049, China
a r t i c l e
i n f o
Article history: Received 5 June 2014 Received in revised form 5 July 2014 Accepted 7 July 2014 Available online 12 July 2014 Keywords: Fluoroquinolones Molecularly imprinted polymers Dummy template Class-selective Matrix solid-phase dispersion.
a b s t r a c t A series of novel dummy molecularly imprinted polymers (DMIPs) were prepared as highly class-selective sorbents for fluoroquinolones. A non-poisonous dummy template, daidzein, was used for the first time to create specific molecular recognition sites for fluoroquinolones in the synthesized polymers. The influence of porogen polarity on dummy molecular imprinting effect was studied. The DMIP prepared using dimethylsulfoxide−acetonitrile (1:1.8, v/v) as porogen achieved the highest imprinting factors (IF) for fluoroquinolones over a range of IF 13.4–84.0. This DMIP was then used for selective extraction of eight fluoroquinolones (fleroxacin, ofloxacin, norfloxacin, pefloxacin, ciprofloxacin, lomefloxacin, enrofloxacin and gatifloxacin) from fish samples based on dummy molecularly imprinted matrix solid-phase dispersion (DMI-MSPD). The extracted fluoroquinolones were subsequently analyzed by high-performance liquid chromatography (HPLC) equipped with a fluorescence detector (FLD). The developed method had acceptable recoveries (64.4−102.7%) and precision (RSDs: 1.7−8.5%, n = 5) for determination of fluoroquinolones in fish samples fortified at levels of 10 and 100 ng g−1 . The limits of detection (LODs) for identification of eight fluoroquinolones ranged between 0.06 and 0.22 ng g−1 . The results demonstrated great potential of the optimized method for sample preparation in routine analysis of trace fluoroquinolones in fish samples. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Fluoroquinolones (FQs) are one of the most important classes of synthetic antibiotics, which show excellent activity against pathogenic Gram-negative and Gram-positive bacteria, as well as anaerobes [1]. The extensive use of fluoroquinolones in foodproducing animals has raised concerns about the potential risks of their residues in foodstuffs of animal origin [2] or environmental matrix [3–5]. These residues can be directly toxic or can elevate the pathogen resistance levels and possible allergic hypersensitivity reactions in humans [6,7]. Maximum residue limits (MRLs) for several FQs have been established in many countries for animal foods [8]. Therefore, monitoring of FQ residues in livestock, poultry, fish and other animal products for human consumption is greatly significant. Currently, the most used methods for fluoroquinolone analysis are based on high-performance liquid chromatography (HPLC),
∗ Corresponding author. Tel.: +86 411 84379562; fax: +86 411 84379562. E-mail address: [email protected] (J. Chen). http://dx.doi.org/10.1016/j.chroma.2014.07.007 0021-9673/© 2014 Elsevier B.V. All rights reserved.
mainly coupled with fluorescence, ultraviolet (UV) or mass spectrometric (MS) detection [9–11]. Different sample preparation methods have been developed for the analysis of fluoroquinolones in biological samples, such as liquid extraction (LE), protein precipitation (PP), liquid–liquid extraction (LLE), solid-phase extraction (SPE), supercritical fluid extraction (SFE) and so on [8,12]. The main limitations of these methods include the relatively low recoveries, the tedious and time-consuming extraction and cleanup processes, the use of a large amount of toxic organic solvents and the low specificity toward fluoroquinolones. Hence alternative sample preparation methods which can tackle these issues are highly desirable for the extraction and enrichment of fluoroquinolones. Matrix solid-phase dispersion (MSPD), first introduced in 1989 by Barker et al. [13], is one of the most promising techniques for the simultaneous disruption, extraction and cleanup of solid, semisolid or viscous samples [14]. MSPD involves blending a viscous, solid or semi-solid sample with an appropriate sorbent (silica, florisil, alumina, C18, C8, etc.) followed by packing, washing and elution. In MSPD, the procedures of homogenization, disruption, extraction and cleanup are combined into one simple process, thus
2
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
greatly reducing analysis time and solvent use and increasing sample throughput. In recent years, MSPD has been widely used in the pre-treatment of food, plant, animal tissues, human biological samples, environmental samples and cosmetics [15]. However, due to the lack of special selectivity, MSPD using traditional sorbents was confronted with the difficulty of selectively extracting target analytes from complex samples. Molecular imprinting is a rapidly developing technique for preparation of polymers with excellent molecular recognition properties, which has been found effective in a variety of extraction techniques such as molecular imprinted solid-phase extraction (MI-SPE), molecular imprinted solid-phase microextraction (MISPME), molecular imprinted stir-bar sorptive extraction (MI-SBSE), molecular imprinted matrix solid-phase dispersion (MI-MSPD) [16] and so on. Fluoroquinolone-imprinted polymers have been applied for their enrichment and purification from complex samples such as milk [17], eggs [18], urine [19], fish [20], serum [21], water [22,23] and soil [24]. Good recoveries as well as high selectivity have been obtained due to the good molecular recognition properties of molecularly imprinted polymers (MIPs). However, all the FQs-MIPs reported previously were prepared using one (pefloxacin [25], gatifloxacin [26], ciprofloxacin, ofloxacin [27] and enrofloxacin) or two (levofloxacin−ciprofloxacin [28]) of fluoroquinolones as templates, where possible leakage of template molecules still happened even after exhaustive washing steps. Template leakage could have a serious impact on the accuracy of analytical method [29] or made it unsuitable for simultaneous analysis of the whole class of fluoroquinolones. This problem has become one of the major areas of concern in sample pretreatment methods employing MIPs [30]. The use of a dummy molecule presents an easy solution to circumvent this problem since any leakage will be different from the analyte [31]. Until now, there has been no report about the use of dummy template that does not belong to fluoroquinolones for the preparation of dummy molecularly imprinted polymers (DMIPs) for fluoroquinolones. This work presents the first attempt of using daidzein as a non-poisonous dummy template for the imprinting of fluoroquinolones. The synthesized DMIPs were used as selective MSPD sorbents for simultaneous determination of eight fluoroquinolones in fish samples. The selectivity, recovery and precision of the developed method were also evaluated.
2. Experimental 2.1. Chemicals and reagents Daidzein was obtained from Zhongxin Pharmaceuticals (Tianjin, China). Enoxacin (ENO), norfloxacin (NOR), ciprofloxacin (CIP), pefloxacin (PEFX), ofloxacin (OFL), lomefloxacin (LOM), fleroxacin (FLX), enrofloxacin (ENR) and gatifloxacin (GAT) were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). The structures of these fluoroquinolones and daidzein are shown in Fig. 1. Diethylstilbestrol (DES), ethylene dimethacrylate (EDMA) and trifluoroacetic acid (TFA) were purchased from J&K Chemical Ltd. (Beijing, China). The initiator 2,2 -azobisisobutyronitrile (AIBN) was supplied by Aladdin Chemical (Shanghai, China). 4-Vinylpyridine (4-VP) was obtained from Acros (NJ, USA). HPLC-grade acetonitrile and methanol were purchased from Fisher (Fair Lawn, NJ, USA). Dimethylsulfoxide (DMSO) and tetrahydrofuran (THF) were purchased from Tianjin Bodi Chemical Engineering Co., Ltd. (Tianjin, China). Dimethylformamide (DMF) was obtained from Tianjin Damao Chemical Reagent Factory (Tianjin, China). Purified water by Milli-Q Plus water purification system (Millipore, Bedford, with a 70 mm × 0.22 m filter) was used throughout the experiments.
2.2. Preparation of imprinted and non-imprinted polymers The dummy molecularly imprinted polymers were prepared using different porogens. The composition of the prepolymerization mixtures are presented in Table 1. The dummy template daidzein (0.2542 g, 1 mmol) was dissolved in 5.6 mL of porogen in a 10 mL thick-walled glass tube. The functional monomer 4-VP (0.42 mL, 4 mmol), cross-linking monomer (EGDMA) (3.8 mL, 20 mmol) and the initiator (AIBN) (0.04 g) were then successively added to the above solution. The solution was sonicated and saturated with dry nitrogen for 10 min before the glass tube was sealed. The tube was then placed in a water bath at 60 ◦ C for 24 h. The obtained DMIPs were crushed, ground, sieved (particle size range: 38.5–63.0 m), sedimented in acetone and dried under vacuum. The template was extracted by extensive washing with methanol−acetic acid (9:1, v/v). The non-imprinted
Fig. 1. Chemical structures of the dummy template, structure analogue and selected fluoroquinolones.
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7 Table 1 Composition of the polymerization mixtures used for the preparation of the dummy molecularly imprinted polymers (DMIPs) and corresponding non-imprinted polymers (NIPs). Polymer DMIP1 NIP1 DMIP2 NIP2 DMIP3 NIP3 DMIP4 NIP4
Daidzein (mmol)
4-VP (mmol)
EGDMA (mmol)
Porogen
1 — 1 — 1 — 1 —
4 4 4 4 4 4 4 4
20 20 20 20 20 20 20 20
DMSO DMSO DMSO:ACN (1:1.8, v/v) DMSO:ACN (1:1.8, v/v) DMF:ACN (1:1.8, v/v) DMF:ACN (1:1.8, v/v) DMF:THF (1:1.8, v/v) DMF:THF (1:1.8, v/v)
polymers (NIP) were prepared simultaneously using the same protocol in the absence of the template molecule. 2.3. Chromatographic evaluation of the polymers The DMIPs and NIPs were suspended in methanol, sonicated in a water bath and then slurry packed into stainless steel HPLC columns (100 × 4.6 mm i.d.) at 3000 psi using an air-driven fluid pump (Haskel, Burbank, CA, USA) with ethanol as the pushing solvent. The chromatographic evaluation of the polymers was carried out using a Waters 515 HPLC pump equipped with a Waters 2487 dual absorbance detector. The binding affinities of polymers toward dummy template were evaluated with acetonitrile as mobile phase and daidzein as analyte. Then, the cross-selectivity of the polymers toward fluoroquinolones was investigated using methanol as mobile phases and enoxacin, fleroxacin, ofloxacin, norfloxacin, pefloxacin, ciprofloxacin, lomefloxacin, enrofloxacin, gatifloxacin, diethylstilbestrol and daidzein as analytes. The results were presented and discussed in Sections 3.1.1 and 3.1.2, respectively. A 20 L aliquot of the analyte (20 ppm) was injected for the analysis with a flow rate of 1 mL min−1 . The UV detector was set at 280 nm. Acetone was injected as a void marker. The capacity factor (k) was calculated as k = (tR − t0 )/t0 , where tR and t0 are the retention times of the analyte and the void marker, respectively. The molecular imprinting factor (IF) was calculated as IF = kMIP /kNIP , where kMIP and kNIP are the capacity factors of the analyte on the imprinted and non-imprinted polymers, respectively. 2.4. MSPD procedure The fish samples (about 500 g) used for this study were collected from a local market. The fish muscle (edible parts) was minced with a meat grinder and stored in well-sealed containers at −20 ◦ C before analysis. One hundred and fifty milligrams of DMIP2 was placed in a glass mortar, and 200 mg of fish meat was added onto the DMIP2. An appropriate volume of fluoroquinolone standards was injected randomly into the fish meat and allowed to stand for 60 min at room temperature. Blank samples were prepared similarly, except that the same volume of methanol containing no fluoroquinolone was added. After adding 200 L of water, the fish meat was blended with the DMIP2 material using a glass pestle until a homogeneous mixture was achieved (∼5 min). The homogenized sample was introduced into an SPE cartridge packed with 50 mg of DMIP2 (equilibrated with 2.0 mL of acetonitrile and 2.0 mL of water). The column was then tapped slightly to remove any air pocket. After rinsing with 3.0 mL of methanol−water (20:80, v/v), the fluoroquinolones were eluted with 4.0 mL of acetonitrile−trifluoroacetic acid (99:1, v/v). The eluents were dried under a moderate stream of nitrogen gas and then reconstituted in 500 L of methanol−water (50:50, v/v). An aliquot of 20 L was injected into the HPLC system for analysis.
3
Table 2 The dielectric constant (ε), Snyder polarity index (SPI), the capacity factors (k) and imprinting factors (IF) of DMIP1, DMIP2, DMIP3 and DMIP4 towards dummy template (daidzein). Polymer
Porogen
εa
SPIb
kMIP
kNIP
IF
DMIP1 DMIP2 DMIP3 DMIP4
DMSO DMSO:ACN (1:1.8, v/v) DMF:ACN (1:1.8, v/v) DMF:THF (1:1.8, v/v)
46.70 39.37 36.89 18.90
7.20 6.20 5.96 4.89
9.99 16.56 27.80 33.23
5.22 4.11 5.19 5.22
1.9 4.0 5.4 6.4
ε and SPI values of mixed solvents were calculated by the equation: εmix =
ϕi εi
and SPImix = ϕi SPIi . Coefficients ϕi represent the mole fraction of solvent. The ε and SPI values of pure DMSO, DMF, ACN and THF were getting from the literatures [32–34]. a Measured dielectric constant. b Snyder polarity index [35].
2.5. HPLC analysis HPLC analysis was performed on an Agilent 1200 system (Agilent Technologies, Palo Alto, CA, USA) equipped with a vacuum degasser, a quaternary pump, an auto-sampler and a fluorescence detector (FLD) connected to a reversed-phase column (Agilent ZORBAX SB-C18, 250 × 4.6 mm i.d., particle size, 5 m). A gradient program was used at a flow rate of 1 mL min−1 , by combining solvent A of water−formic acid (99.9:0.1, v/v), solvent B of methanol and solvent C of acetonitrile as follows: 9.5−11.5% C (24 min), 11.5−26.5% C (10 min), 26.5−96.5% C (2 min), 96.5% C (5 min). Solvent B was maintained at 3.5% during the whole separation process. The column temperature was kept at 15 ◦ C. The injection volume was 20 L, and the FLD excitation and emission wavelengths were set at 290 and 480 nm, respectively. 3. Results and discussion 3.1. Synthesis and chromatographic evaluation of the DMIPs The molecularly imprinted polymers for fluoroquinolones (FQs) were prepared by using daidzein as the dummy template to avoid the template bleeding problem. Daidzein can be dissolved well only in organic solvents such as DMSO and DMF. In order to improve the imprinting effect, less polar solvents such as ACN and THF were added into DMSO or DMF up to a maximum proportion while still maintaining good solubility for the dummy template. The binary mixed solvents with different polarities were used as porogens for the polymerization and their effects on the affinity and selevtivity of the synthesised DMIPs toward dadzein and FQs were investigated. 3.1.1. Effect of porogen polarity on the selectivity of DMIPs toward daidzein The DMIPs were first evaluated in terms of their capacity factors (kMIP ) and imprinting factors (IF) toward the dummy template molecule, daidzein. ACN was used as the mobile phase. The polarity of porogen solvent is described by both Snyder polarity index (SPI) and dielectric constant (ε). The relationship between porogen polarities and the imprinting factors of daidzein on DMIPs (DMIP1, DMIP2, DMIP3 and DMIP4) was discussed. As can be seen from Table 2, DMIP4 synthesized from the porogen with the lowest SPI and ε values achieved the highest capacity and imprinting factor for daidzein. Negative correlations between the SPI/ε values and the IF values of DMIPs were obtained by further correlation analysis and linear fitting, as shown in Fig. 2. SPI was found to be a more adequate parameter to describe the porogen polarity with a square value of correlation coefficient (r2 ) of 0.93. These results can be associated with the stronger interactions between basic -donor/acceptor monomer (4-VP) and daidzein in
4
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
Fig. 2. Influence of dielectric constant (ε) and Snyder polarity index (SPI) of the porogens on the imprinting factors (IF) of DMIPs towards daidzein.
less polar solvents, thus creating more high-affinity binding sites during the polymerization [36]. 3.1.2. Effect of porogen polarity on the cross-selectivity of DMIPs toward fluoroquinolones The dummy molecular imprinting effects of DMIPs toward fluoroquinolones were investigated by using acetonitrile, methanol, water or their mixture as the mobile phases. The best selectivity for fluoroquinolones was found when methanol was used, and the resulted kNIP , kMIP and IF values are listed in Fig. 3. As can be seen in Fig. 3A–D, DMIP1 and DMIP2 showed higher capacities and imprinting factors for fluoroquinolones as compared to
DMIP3 and DMIP4. This observation was significantly different from the case for daidzein. The negative correlations between SPI/ε and the imprinting factors for daidzein no longer exist for fluoroquinolones. The most plausible reason of these results is the different cavity sizes produced from daidzein template by using porogens with different polarities. Porogens with lower polarities can lead to the formation of more rigid cavities for daidzein, which are not easily accessible to larger fluoroquinolone molecules. Therefore, the strong acid–base electrostatic interactions between 4-VP and fluoroquinolones in the cavities are disabled, resulting in very low affinities. DMIP2 with suitable imprinted cavity size by using dimethylsulfoxide−acetonitrile (1:1.8, v/v) as the porogen achieved the highest imprinting factors for fluoroquinolones over a range of 13.4–84.0. Furthermore, we can see from Fig. 3B that the capacity factors of different fluoroquinolones on DMIP2 were inversely related to their molecular sizes. The smallest ENO and NOR molecules among the fluoroquinolones obtained the highest capacity factors. The methyl groups of OFL, LOM and PEFX on piperazine rings situated at position 7 reduced their capacity factors to some degree due to their bulky molecules. In addition, the large ethyl fluoride or cyclopropyl ring groups of FLX, CIP, ENR and GAT dramatically reduced their retention abilities on the DMIP2. In summary, porogen polarity plays a very important role in the affinity and selectivity of dummy molecular imprinting. Highly efficient imprinting of dummy template does not always result in high affinities of target molecules when the structures of dummy template and target molecules are not exactly the same. The recognition abilities of DMIP2 for fluoroquinolones presented here were excellent and the IF values tested were much higher than those reported in literatures [23,26]. Finally, low affinities were observed
Fig. 3. Evaluation of the cross-selectivity of DMIPs in terms of capacity factors (k) and imprinting factors (IF) towards ENO, FLX, OFL, NOR, PEFX, CIP, LOM, ENR, GAT, DAI and DES: (A) DMIP1, (B) DMIP2, (C) DMIP3, (D) DMIP4. Conditions: column, 150 mm × 4.6 mm; mobile phase, methanol; flow rate, 1 mL min−1 ; sample volume, 20 L (20 ppm); UV detection, 280 nm (see Section 2.3). Acetone was used as the void marker.
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
5
Fig. 4. MSPD extraction recoveries (%) obtained with the DMIP2 and the NIP sorbents for eight fluoroquinolones from fish samples spiked at 100 ng g−1 level using a washing step with 3 mL of methanol−water (20:80, v/v).
for daidzein and DES when methanol was used as the mobile phase with imprinting factors of 1.08 and 1.29, respectively, which further confirmed the good class selectivity of DMIP2 for fluoroquinolones. 3.2. Optimization of DMI-MSPD procedure 3.2.1. Effect of the mass ratio of sample to DMIP2 The first step for the development of an MSPD method was selecting a suitable mass ratio of sample/sorbent, in order to allow complete adsorption of the sample components and to facilitate the sample transfer onto the cartridge. The sample/sorbent ratios typically ranged from 1:1 to 1:4 using traditional SPE sorbents in the previously reported works [37]. Another research reported the optimal sample/sorbent ratio of 1:1 using FQs-MIPs as sorbents [18]. In our study, different sample/sorbent ratios were evaluated (2:1, 4:3, 1:1 and 2:3) using 200 mg of the sample. The results showed that a sample/sorbent ratio of 4:3 was sufficient for complete retention of fluoroquinolones, with recoveries ranging from 66.8 to 104.2% for triplicate experiments. Further increasing the proportion of sorbents gave no improvement for the recoveries of fluoroquinolones. However, less DMIP2 sorbents resulted in heterogeneous sample mixture and low recoveries (e.g. 48.2−86.7% for 2:1). Therefore, the sample/sorbent ratio of 4:3 was finally selected for further investigations. 3.2.2. Effect of washing and eluting conditions The nature and volume of washing and elution solvent are the key factors, which should be carefully optimized to achieve high recoveries for analytes, while eliminating most of the interferences originating from a complex biological matrix. The effect of washing solvent type and volume on recoveries and selectivity was studied based on 150 mg of DMIP2 sorbent and 200 mg of fish sample (100 ng g−1 of spiking level for each fluoroquinolone). According to Section 3.1.2, high selectivity of DMIP2 for fluoroquinolones can be achieved by using methanol as the mobile phase. Therefore, methanol was used as the selective washing solvent at first. Unfortunately, despite its high selectivity, the use of methanol led to low recoveries due to its strong elution strength. When 2 mL of methanol was used, the recoveries less than 40% were obtained on DMIP2. The mixtures of water−methanol (methanol content ranging from 0 to 50%) were tested. High methanol content resulted in low recoveries of fluoroquinolones. When 2 mL of water−methanol (50:50, v/v) was used, the recoveries from 36.4 to 84.8% were obtained. In order to achieve higher recoveries, water−methanol (80:20, v/v) was finally selected. Different volumes of water−methanol (80:20, v/v) were investigated from 2.0 to 4.0 mL, and 3.0 mL was found to be the optimum volume of washing solvent. The results are shown in Fig. 4.
Fig. 5. Chromatograms of spiked fish samples (10 ng g−1 ). Blue line directly extracted fish sample (200 mg of fish sample extracted by 4 mL of acetonitrile−TFA (99:1, v/v) in an ultrasonic bath for 5 min); red line fish sample after DMI-MSPD process. Peaks: 1, FLX; 2, OFL; 3, NOR; 4, PEFX; 5, CIP; 6, LOM; 7, ENR; 8, GAT.
In this case, the extraction recoveries for eight fluoroquinolones on the NIP ranged between 32.4 and 86.1%, whereas those on DMIP2 ranged between 66.8 and 104.2%. These results further confirmed the existence of the specific interactions between fluoroquinolones and DMIP2. The low selectivity of DMIP2 for ENR and PEFX was mostly due to the high hydrophobic interactions (nonspecific) resulted from their relatively high octanol/water partition coefficients (log Kow ). The log Kow values are 0.22 and 0.31 for ENR and PEFX, respectively, while those of the other fluoroquinolones are all below zero [38]. The final elution of fluoroquinolones was conducted by using methanol−ammonia (99:1, v/v), methanol−TFA (99:1, v/v) and acetonitrile−TFA (99:1, v/v). The best recoveries were obtained by using acetonitrile−TFA (99:1, v/v) as elution solvent with a relatively clean baseline. 3.3. Method validation Two calibration curves were established for each fluoroquinolone to cover a wide concentration range 2.5−400 ng g−1 , i.e. a low curve for concentrations below 30 ng g−1 and a high curve for concentrations above 30 ng g−1 . Good linearity was obtained for all analytes in both low and high concentrations, with regression coefficients R > 0.999 (Table 3). To evaluate the applicability of the optimized DMI-MSPD-HPLCFLD procedure to real samples, accuracy and precision of the method were evaluated using spiked blank fish samples at two concentration levels (10 and 100 ng g−1 ). No background of FQs was found in the blank fish samples. The results summarized in Table 3 show acceptable recoveries, repeatability and intermediate precision of the method. For repeatability, five extractions were performed on the same day under the optimum conditions, and the RSD values ranged from 1.7 to 5.7%. The intermediate precision was determined by measuring the RSDs of experiments carried out in different days (n = 5), and day-to-day precision was between 2.4 and 8.5%. The average recoveries of fluoroquinolones at two different levels were ranged from 64.4 to 102.7%. Moreover, as shown in Fig. 5, the use of DMIP2 resulted in less matrix interferences compared to the directly extracted sample, and none of co-elution peaks was observed in the chromatogram. The recoveries obtained for all the fluoroquinolones except LOM were excellent and better than those in literatures using other MIP materials or traditional SPE sorbents [19,20,39]. The limits of detection (LODs) of the DMIMSPD method, calculated as a signal-to-noise ratio of 3, were found
6
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7
Table 3 Correlation coefficients, repeatability, intermediate precision, limits of detection and recovery study of the proposed method. Analyte
FXL OFL NOR PEFX CIP LOM ENR GAT a b
Correlation coefficient R2
Repeatability RSD (%, n = 5)
Low curvea
High curveb
10 ng g−1
100 ng g−1
10 ng g−1
100 ng g−1
0.9999 0.9997 0.9991 0.9996 0.9999 0.9999 0.9999 0.9999
0.9996 0.9994 0.9998 0.9993 0.9999 0.9996 0.9993 0.9998
2.0 5.0 3.8 4.1 5.7 2.0 3.9 4.8
1.7 3.2 5.3 2.7 2.7 5.0 2.5 2.6
3.5 7.7 2.6 2.6 2.4 7.9 8.5 6.3
5.5 6.5 5.1 5.7 7.0 5.2 6.5 3.9
Intermediate precision RSD (%, n = 5)
LOD S/N = 3 (ng g−1 )
0.16 0.20 0.14 0.22 0.21 0.06 0.06 0.07
Average recoveries (%) 10 ng g−1
100 ng g−1
93.1 96.7 99.2 94.3 102.7 88.4 99.1 100.0
89.8 98.0 95.0 99.4 96.4 64.4 95.2 90.2
The calibration curve for low concentration levels (2.5, 5, 10, 20, 30 ng g−1 ). The calibration curve for high concentration levels (30, 50, 100, 200, 400 ng g−1 ).
to be in the range 0.06−0.22 ng g−1 (Table 3). The values obtained are lower than those mentioned in previous papers [19,20,39] and are well below the maximum residue levels (MRLs) set by the European Union [8]. The good recoveries, low LODs and excellent accuracy demonstrated the applicability of the DMIP sorbent for pre-treatment of fluoroquinolones in fish samples. 4. Conclusions In this work, highly class-selective dummy molecularly imprinted polymers (DMIPs) for fluoroquinolones were successfully synthesized by using a non-poisonous dummy template molecule (daidzein). The porogen polarity has a significant influence on the dummy molecular imprinting effect. Among the polymers prepared, the DMIP prepared using dimethylsulfoxide−acetonitrile (1:1.8, v/v) as porogen showed the highest imprinting factors for eight fluoroquinolones. A simple and highly selective sample preparation method based on the DMIP for simultaneous determination of eight fluoroquinolones in fish samples was developed by DMIP-based MSPD followed by HPLC-FLD. The new DMI-MSPD-HPLC-FLD method was not only highly selective and accurate, but it also effectively eliminated the template bleeding problem. These features demonstrate great potential of the optimized method for selective extraction of fluoroquinolones from fish samples at the ng g−1 level. Acknowledgments This work was supported by the National Natural Science Foundation of China (Nos. 21277139, 21205117) and National High Technology Research and Development Program of China (863 program, No. 2013AA065203). References [1] S.A. Brown, Fluoroquinolones in animal health, J Vet Pharmacol Ther 19 (1996) 1–14. [2] M.D. Barton, Antibiotic use in animal feed and its impact on human health, Nutr Res Rev 13 (2000) 279–300. [3] Y. Pico, V. Andreu, Fluoroquinolones in soil—risks and challenges, Anal Bioanal Chem 387 (2007) 1287–1299. [4] H. Sanderson, D.J. Johnson, T. Reitsma, R.A. Brain, C.J. Wilson, K.R. Solomon, Ranking and prioritization of environmental risks of pharmaceuticals in surface waters, Regul Toxicol Pharm 39 (2004) 158–183. [5] P. Sukul, M. Spiteller, Fluoroquinolone antibiotics in the environment, Rev Environ Contam Toxicol 191 (2007) 131–162. [6] J. Davies, D. Davies, Origins and evolution of antibiotic resistance, Microbiol Mol Biol Rev 74 (2010) 417–433. [7] A.D. Anderson, J.M. Nelson, S. Rossiter, F.J. Angulo, Public health consequences of use of antimicrobial agents in food animals in the United States, Microb Drug Resist 9 (2003) 373–379. [8] V. Andreu, C. Blasco, Y. Picó, Analytical strategies to determine quinolone residues in food and the environment, Trends Anal Chem 26 (2007) 534–556.
[9] M. Farré, D. Barceló, Analysis of emerging contaminants in food, Trends Anal Chem 43 (2013) 240–253. [10] M. Seifrtova, L. Novakova, C. Lino, A. Pena, P. Solich, An overview of analytical methodologies for the determination of antibiotics in environmental waters, Anal Chim Acta 649 (2009) 158–179. [11] J. Sousa, G. Alves, J. Abrantes, A. Fortuna, A. Falcao, Analytical methods for determination of new fluoroquinolones in biological matrices and pharmaceutical formulations by liquid chromatography: a review, Anal Bioanal Chem 403 (2012) 93–129. [12] J.Y. Shen, M.R. Kim, C.J. Lee, I.S. Kim, K.B. Lee, J.H. Shim, Supercritical fluid extraction of the fluoroquinolones norfloxacin and ofloxacin from orally treated-chicken breast muscles, Anal Chim Acta 513 (2004) 451–455. [13] S.A. Barker, A.R. Long, C.R. Short, Isolation of drug residues from tissues by solid phase dispersion, J Chromatogr A 475 (1989) 353–361. [14] S.A. Barker, Matrix solid-phase dispersion, J Chromatogr A 885 (2000) 115–127. [15] A.L. Capriotti, C. Cavaliere, A. Laganà, S. Piovesana, R. Samperi, Recent trends in matrix solid-phase dispersion, Trends Anal Chem 43 (2013) 53–66. [16] E. Turiel, A. Martin-Esteban, Molecularly imprinted polymers for sample preparation: a review, Anal Chim Acta 668 (2010) 87–99. [17] Y. Lv, L. Yang, X. Liu, Z. Guo, H. Sun, Preparation and evaluation of a novel molecularly imprinted hybrid composite monolithic column for online solidphase extraction coupled with HPLC to detect trace fluoroquinolone residues in milk, Anal Methods 5 (2013) 1848–1855. [18] H. Yan, F. Qiao, K.H. Row, Molecularly imprinted-matrix solid-phase dispersion for selective extraction of five fluoroquinolones in eggs and tissue, Anal Chem 79 (2007) 8242–8248. [19] E. Benito-Pena, S. Martins, G. Orellana, M.C. Moreno-Bondi, Water-compatible molecularly imprinted polymer for the selective recognition of fluoroquinolone antibiotics in biological samples, Anal Bioanal Chem 393 (2009) 235–245. [20] W. Junping, P. Mingfei, F. Guozhen, W. Shuo, Preparation of a novel molecularly imprinted polymer by a sol-gel process for on-line solid-phase extraction coupled with high performance liquid chromatography to detect trace enrofloxacin in fish and chicken samples, Microchim Acta 166 (2009) 295–302. [21] D. Xiao, P. Dramou, N. Xiong, H. He, H. Li, D. Yuan, H. Dai, Development of novel molecularly imprinted magnetic solid-phase extraction materials based on magnetic carbon nanotubes and their application for the determination of gatifloxacin in serum samples coupled with high performance liquid chromatography, J Chromatogr A 1274 (2013) 44–53. [22] A. Prieto, S. Schrader, C. Bauer, M. Moeder, Synthesis of a molecularly imprinted polymer and its application for microextraction by packed sorbent for the determination of fluoroquinolone related compounds in water, Anal Chim Acta 685 (2011) 146–152. [23] E. Benito-Pena, J.L. Urraca, B. Sellergren, M. Cruz Moreno-Bondi, Solid-phase extraction of fluoroquinolones from aqueous samples using a watercompatible stochiometrically imprinted polymer, J Chromatogr A 1208 (2008) 62–70. [24] E. Turiel, A. Martin-Esteban, J.L. Tadeo, Molecular imprinting-based separation methods for selective analysis of fluoroquinolones in soils, J Chromatogr A 1172 (2007) 97–104. [25] M. Zheng, R. Gong, X. Zhao, Y. Feng, Selective sample pretreatment by molecularly imprinted polymer monolith for the analysis of fluoroquinolones from milk samples, J Chromatogr A 1217 (2010) 2075–2081. [26] D. Xiao, P. Dramou, N. Xiong, H. He, D. Yuan, H. Dai, H. Li, X. He, J. Peng, N. Li, Preparation of molecularly imprinted polymers on the surface of magnetic carbon nanotubes with a pseudo template for rapid simultaneous extraction of four fluoroquinolones in egg samples, Analyst 138 (2013) 3287–3296. [27] F. Qiao, H. Sun, Simultaneous extraction of enrofloxacin and ciprofloxacin from chicken tissue by molecularly imprinted matrix solid-phase dispersion, J Pharm Biomed Anal 53 (2010) 795–798. [28] P. Liu, J. Shen, L. Gao, L. Liu, R. Li, Q. Li, Determination of fluoroquinolones in milk by high-performance liquid chromatography using mixed-templates imprinted polymer extraction, Asian J Chem 22 (2010) 6275–6288.
X. Sun et al. / J. Chromatogr. A 1359 (2014) 1–7 [29] F.G. Tamayo, E. Turiel, A. Martin-Esteban, Molecularly imprinted polymers for solid-phase extraction and solid-phase microextraction: recent developments and future trends, J Chromatogr A 1152 (2007) 32–40. [30] L. Chen, S. Xu, J. Li, Recent advances in molecular imprinting technology: current status, challenges and highlighted applications, Chem Soc Rev 40 (2011) 2922–2942. [31] L. Andersson, A. Paprica, T. Arvidsson, A highly selective solid phase extraction sorbent for pre-concentration of sameridine made by molecular imprinting, Chromatographia 46 (1997) 57–62. [32] P.B. Wright, A.S. Lister, J.G. Dorsey, Behavior and use of nonaqueous media without supporting electrolyte in capillary electrophoresis and capillary electrochromatography, Anal Chem 69 (1977) 3251–3259. [33] T. Liu, I.A. Nicholls, T. Oberg, Comparison of theoretical and experimental models for characterizing solvent properties using reversed phase liquid chromatography, Anal Chim Acta 702 (2011) 37–44.
7
[34] X. Li, R. Chen, M. Huang, M. Zhu, Q. Chen, Synthesis of a soluble pyrrole copolymer with phenetidine, J Polym Sci A Polym Chem 42 (2004) 2073–2092. [35] L. Snyder, Classification off the solvent properties of common liquids, J Chromatogr Sci 16 (1978) 223–234. [36] S. Pardeshi, R. Dhodapkar, A. Kumar, Influence of porogens on the specific recognition of molecularly imprinted poly(acrylamide-co-ethylene glycol dimethacrylate), Compos Interfaces 21 (2014) 13–30. [37] E.M. Kristenson, U.A.T. Brinkman, L. Ramos, Recent advances in matrix solidphase dispersion, Trends Anal Chem 25 (2006) 96–111. [38] D.L. Ross, C.M. Riley, Aqueous solubilities of some variously substituted quinolone antimicrobials, Int J Pharm 63 (1990) 237–250. ˜ ˜ Simultaneous [39] A. Espinosa-Mansilla, A. Jiménez Girón, A. Munoz de la Pena, determination of the residues of fourteen quinolones and fluoroquinolones in fish samples using liquid chromatography with photometric and fluorescence detection, Czech J Food Sci 30 (2012) 74–82.