Capillary electrophoresis–chemiluminescence determination of norfloxacin and prulifloxacin

a n a l y t i c a c h i m i c a a c t a 6 2 3 ( 2 0 0 8 ) 231–237

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/aca

Capillary electrophoresis–chemiluminescence determination of norfloxacin and prulifloxacin Zhongju Yang, Xiaoli Wang, Weidong Qin ∗ , Huichun Zhao ∗∗ College of Chemistry, Beijing Normal University, Beijing 100875, PR China

a r t i c l e

i n f o

a b s t r a c t

Article history:

A capillary electrophoresis (CE)–chemiluminescence (CL) method for determining nor-

Received 4 April 2008

floxacin (NFLX) and prulifloxacin (PFLX) was developed based on the enhanced CL intensity

Received in revised form 8 June 2008

of the cerium(IV)–sulfite–fluoroquinolone (FQ) reaction sensitized by terbium(III). The sep-

Accepted 14 June 2008

aration was conducted in buffer composed of 20 mM sodium citrate, 4 mM citric acid and

Published on line 24 June 2008

10 mM sodium sulfite at pH 6.1. The CL reagent solution consisted of 2 mM cerium(IV), 4 mM terbium(III) and 1.1 mM hydrochloric acid. NFLX and PFLX were baseline separated within

Keywords:

11 min with detection limits (S/N = 3) of 0.057 and 0.084 ␮g mL−1 , respectively. The maximum

Capillary electrophoresis

intra- and inter-day relative standard deviations (R.S.D.s) of migration time of the analytes

Chemiluminescence

were less than 4.0% and 4.2%, respectively. The proposed method was applied to detect NFLX

Fluoroquinolones

and PFLX in fortified urine sample and the results were comparable to high-performance

Urine sample

liquid chromatography (HPLC)–UV method. Moreover, the high selectivity of the CL detection and the high-separation efficiency of CE render the method the potential of quick analyzing fluoroquinolones in real complex matrix. © 2008 Elsevier B.V. All rights reserved.

1.

Introduction

Fluoroquinolones (FQs) are derivatives of quinolones, obtained by addition of a fluorine atom in position 6 and a piperazine substituent in position 7 [1]. They represent important therapeutic advantages of higher antibacterial activity due to addition of the two groups [2–4]. They inhibit bacterial DNA-gyrase in cells so that bacteria cannot reproduce and dead finally [5]. These compounds have been used extensively in veterinary, clinical medicine and particularly in food-producing animal husbandry. Thus, their elimination in waste products and subsequent presence in the environment and foodstuffs is a concern. The European Union (EU) has made official controls on products of animal origin intended for human consumption and set tolerance levels for these compounds as maximum residue limits (MRLs) in Council

∗

Regulation number 2377/90 and later modifications [6,7]. Norfloxacin (NFLX, Fig. 1a) is one of the third generation members of FQs that is often used for prevention of infections in clinical treatment; while prulifloxacin (PFLX, Fig. 1b) is a new FQ of fourth generation with a potent and broadspectrum antibacterial activity. Although they are not listed in the Regulation, determination of NFLX and PFLX should be also important due to their increasing usage and influence on human body. Capillary electrophoresis (CE) has developed into an attractive analytical tool for the separation of various types of analytes, owing to its high resolution, rapidity, and low sample and electrolyte consumption [8]. CE is nowadays an alterative to high-performance liquid chromatography (HPLC) in drug analysis [9,10]. CE separation of quinolones (including FQs) has been theoretically studied [11,12] and its applica-

Corresponding author at: College of Chemistry, Beijing Normal University, No. 19, Xin Jie Kou Wai Street, Hai Dian District, Beijing 100875, PR China. Tel.: +86 10 58802531; fax: +86 10 58802075. ∗∗ Corresponding author. Fax: +86 10 58802075. E-mail addresses: [email protected] (W. Qin), [email protected] (H. Zhao). 0003-2670/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2008.06.023

232

a n a l y t i c a c h i m i c a a c t a 6 2 3 ( 2 0 0 8 ) 231–237

Fig. 1 – Chemical structures of FQs studied: (a) norfloxacin and (b) prulifloxacin.

tion in assaying quinolone antibiotics and their residues in various matrix, such as rat liver perfusate, plasma, chicken tissues, milk, waste water, and human urine [13–19], has been actively investigated. Because of the low loadability of CE, developing sensitive detection methods is a concern. Faria et al. separated and detected five quinolones employing capillary zone electrophoresis (CZE)–UV technique [20]. In Wang’s work, quinolones were determined by CZE coupled ¨ with end-column amperometric detection [21]. Horstkotter et al. employed CE-laser-induced fluorescence (LIF) in determining quinolone residues in chicken muscle [22]. Fan et al. reported a CE-potential gradient detection method for determining quinolones in milk [15]. Deng et al. reported their work on determining one FQ, norfloxacin, by capillary electrophoresis–electrochemiluminescence [17]. Mass spectrometric detection was also hyphenated to CE in determination of quinolone antibiotics [23–25]. Chemiluminescence (CL), which has been widely used in chemical and biological fields during the past several decades, relies on the effects related to the chemical reaction only, i.e. without the need for an external energy supply [26,27]. It is characterized by minimum background and no light scattering, therefore can offer high sensitivity and the possibility of simple and quick analysis [26,28]. Since Abbott et al. reported the CL determination of morphine and its derivative by oxidation with potassium permanganate [29], there have been increasing interests in the application of CL reactions in analytical chemistry and numerous CL reagents have been assessed using flow-injection (FI) technique [30–32] and CE [33–36]. In recent years, Ce(IV)–sulfite system was actively studied [37,38]. Ce(IV) oxidizes Na2 SO3 and produces excited sulphur dioxide (SO2 *) in acidic solutions [39]. SO2 * emits very low CL because of the low-luminescence efficiency; however, the CL emission will increase dramatically with addition of Tb(III) and FQs [28,40,41]. Studies [40,41] show that Tb(III) and FQs can form complex, which will accept the energy emitted from SO2 *. Tb(III) is then excited due to the intra-molecule energy transferring process, and fluorescence will be generated when it returns to ground state. The method has been employed in conjunction with FI [38] in sensitive detection of quinolones due to the strong CL intensity. However, application of Ce(IV)–sulfite–Tb(III) system in CE–CL was rarely studied. The pI values of FQs studied in this experiment are in the range of 7–8.5 [21] and nearly neutral or alkaline solutions are usually required as background electrolytes in CE for high-separation efficiencies; whereas acidic environment is indispensable to facilitate the CL reaction. Therefore, it is a challenge to control the factors relating to separation and detection. To the best of our knowledge, this is the first report on CE–CL determination of NFLX and PFLX based on

Ce(IV)–sulfite–Tb(III) system. The parameters influencing separation and detection were studied, and the method developed was applied to the determination of NFLX and PFLX in a fortified urine sample. Besides the advantages of rapidity and high resolution, the constituents in the sample that generated peaks in HPLC–UV did not interfere with the peaks of the analytes in CE–CL, suggesting the high potential of the method in determining FQs in complex real samples.

2.

Materials and methods

2.1.

Chemicals, reagents and samples

NFLX and PFLX were purchased from the Medicinal and Biological Research Institute (Beijing, China), and 2-(Ncyclohexylamino) ethanesulfonic acid (CHES) was product of Sigma (St. Louis, MO, USA). Hydrochloric acid, sodium citrate, citric acid, sodium sulfite, imidazole, glycine, phosphoric acid, sodium borate, acetic acid, Tb4 O7 and ammonium cerium nitrate were bought from Beijing Chemical Plant (Beijing, China). Deionized water (Millipore, Milli-Q Water System, Bedford, MA, USA) was used for preparing solutions throughout the experiment. The stock solution of terbium(III) chloride (0.1 M) was prepared by dissolving 0.1888 g Tb4 O7 with 500 ␮L concentrated hydrochloric acid (12.4 M) in a 5-mL beaker and evaporating to dryness with water bath. The residue was dissolved in 10 mL deionized water and stored at 4 ◦ C. Stock solutions of 50 mM Ce(IV) (dissolved in 100 mM nitric acid) and sodium sulfite were prepared daily. All working solutions were prepared by appropriate dilution of the stock solutions and filtered with 0.22-␮m filters (Jiuding High Tech., Beijing, China) before use. The 10,000 ␮g mL−1 individual stock solutions of NFLX and PFLX were prepared by dissolving 0.0100 g standards in 6 mM sodium hydroxide, respectively. Urine samples obtained from health volunteers were diluted 100-fold with deionized water to prevent the chromatographic column damage due to the matrix composition [42] and was spiked with the standards to desired concentrations.

2.2.

Instruments and general procedures

The CE–CL system consisted of a high-voltage power supply (Sanchuan High Tech., Tianjin, China) and a homemade CL detector as shown in Fig. 2. Separation was carried out in a bare fused-silica capillary (Yongnian Optical Fiber, Hebei, China) of 75 ␮m × 47.5 cm. The polyimide coating of 1 cm in length at one end of the separation capillary was burned and removed. The burned-end of the capillary was inserted into a 530-␮m i.d. capillary (Yongnian Optical Fiber), which acted as reaction capillary, to a depth of 1.6 cm. The separation capillary and the reaction capillary were fixed in place by a three-way Plexiglas joint. A detection window of 0.8 cm length was made on the reaction capillary and was placed in front of a photomultiplier tube (PMT, R928, Hamamatsu Photonics, Japan). The outlet of the reaction capillary and the grounding electrode were immersed into the waste vial. The signal from the PMT was collected (at 20 Hz) and processed with HW2000 Chromatography Station (Qianpu, Jiangsu, China). In order to

a n a l y t i c a c h i m i c a a c t a 6 2 3 ( 2 0 0 8 ) 231–237

233

ried out on a 20 cm × 4.6 mm Diamonsil C18 column (5 ␮m, Diamonsil, Beijing, China) at a flow rate of 0.3 mL min−1 .

Fig. 2 – Schematic diagram of CL system: (1) CL reagents; (2) nitrogen gas inlet; (3) separation capillary; (4) CL reagent delivery-capillary; (5) black box; (6) PMT; (7) reaction capillary; (8) grounding electrode; (9) waste vial.

prevent the serious signal distortion stems from the analog circuits [43], the signal from the PMT was transferred directly to the Chromatography Station instead of being filtered with a conventional RC-type low-pass filter. For the purpose of noise reduction that required by quantitative calculations, the data acquired were asynchronously processed with moving average filtering method at a filter width of 19. In order to keep the flow of the reaction solution stable and therefore obtain reproducible CL signal, the CL solution was delivered into the reaction capillary by pressure-driving strategy. The CL solution was added into a bottle, which has an inlet and an outlet on the cap. During work, the inlet was connected to a nitrogen cylinder via a regulator and a pressure was applied to the bottle. The CL solution was driven to the tee via a 25 cm × 75 ␮m capillary. The flow rate was adjusted by varying the pressure applied. During experiment, a 530-␮m i.d. capillary was secured onto a ruler to measure the flow rate. The time needed for the solution passing through a certain length (l = 2 cm in our experiment) of a 530-␮m i.d. measuring capillary was measured with a stopwatch, and the flow rate was calculated using following equation: f =

d2 l 4t

3.

Results and discussion

3.1.

Selection of buffer

The redox reagents influence not only the CL emission intensity but also the separation efficiency. Preliminary experiments showed that addition of cerium(IV) or Tb(III) into buffer resulted in precipitates. High-detection sensitivity could be achieved when sodium sulfite was added into the running buffer and the CL regent of cerium(IV) and Tb(III) was driven into the reaction capillary by pressure. In order to facilitate the negatively charged sulfite migrate to the reaction capillary, the anode of the capillary was lifted above the grounding end. Our experiments revealed that the peak tailing was not significant while the peak heights reached maximum when the anodic end was raised 12 cm against the waste vial. The pKa s of NFLX and PFLX are in the range of 5.10–8.5 [9,45–47]. Buffer of pHs from 5.43 to 9.35, including acetate, acetate–imidazole, glycine–NaOH, CHES–NaOH, borate–phosphate and citrate, were tested. Fig. 3 shows that CL responses from the stan-

(1)

where f is the flow rate; d is the internal diameter of the measuring capillary (530 ␮m); t is the time needed for the solution passing through the l-cm length of the measuring capillary. Variation in pressure of 0.2–0.8 bar corresponded to the flow rate of 5–18 ␮L min−1 . The HPLC system comprised of an Agilent 1100 series (Agilent Tech., Palo Alto, CA, USA) with a manual injection valve fitted with a 20-␮L sample loop. Detection was performed by a Model G1315B UV–vis diode array detector (Agilent) at 254 nm. The HPLC method in Ref. [44] was slightly modified and followed. The mobile phase consisted of acetonitrile and 0.5 M phosphoric acid (20:80, v/v), and the pH of the final mobile phase was adjusted to 3.0 with triethylamine. Elution was car-

Fig. 3 – Comparison of different buffer systems—buffers: (A) 20 mM sodium citrate + 4 mM citric acid + 10 mM sodium sulfite; (B) 10 mM CHES + 5 mM sodium hydroxide + 5 mM sodium sulfite; (C) 10 mM sodium sulfite + 5 mM acetic acid; (D) 30 mM glycine + 5 mM sodium sulfite + 6 mM sodium hydroxide; (E) 10 mM sodium borate + 3 mM phosphoric acid + 5 mM sodium sulfite; (F) 20 mM imidazole + 10 mM acetic acid + 5 mM sodium sulfite. The pH values of the above buffers were (A) 6.1; (B) 9.35; (C) 5.43; (D) 9.11; (E) 9.01 and (F) 6.83. Reagent solution, 2 mM cerium(IV) + 4 mM Tb(III) + 1.1 mM hydrochloric acid; flow rate, 8 ␮L min−1 ; applied voltage, +10 kV; peak identities: (1) NFLX and (2) PFLX. Concentrations of the standards were 10 ␮g mL−1 each.

234

a n a l y t i c a c h i m i c a a c t a 6 2 3 ( 2 0 0 8 ) 231–237

dards are very low in borate–phosphate and acetate–imidazole buffers (traces E and F), although the merged peaks were reproducible and can be distinguished from the baseline as enlarged in the insets. Only one peak could be obtained with glycine–NaOH, CHES–NaOH and acetate systems. Moreover, the detection sensitivity was low with glycine–NaOH buffer. The merged peaks of NFLX and PFLX in CHES–NaOH and acetate buffers could not be resolved by varying the buffer pH. However, both the signal intensity and the resolution could be adjusted and improved with citrate system. Hence, citrate buffer was employed for the subsequent studies. Our experiments showed that the chemiluminescence detector could not give rise to signal corresponding to the electroosmotic flow, as there were no luminescent substances in the deionized water.

3.2.

Influence of buffer pH

NFLX and PFLX molecules are zwitters because they contain nitrogen atom (proton acceptor) and a carboxyl group (proton donor). Variation in buffer pH would change their charge and mobilities [12,15]. The CL intensity of the two FQs is low and the peaks cannot be resolved at pH 5.4 (Fig. 4). They are separated at pH 6.1 and 6.6, while the peak heights are highest at pH 6.1. The migration times of the analytes increase with buffer pH, which is mainly attributable to the increasing negative charge density of the FQ molecules at higher pH. The resolution between the peaks was calculated to be 0, 5.5 and 3.2 under pH value of 5.4, 6.1 and 6.6, respectively. In order for the maximum detection sensitivity and resolution, running buffer consisting of 20 mM sodium citrate, 4 mM citric acid and 10 mM sodium sulfite at pH 6.1 was employed.

3.3.

Effect of cerium in the CL reagent

Fig. 5 shows that highest CL intensity was obtained with 2 mM Ce(IV); further increase in Ce(IV) concentration resulted in decreasing peak heights. Experiments indicated that repeata-

Fig. 5 – Influence of cerium concentration. Buffer: 20 mM sodium citrate + 4 mM citric acid + 10 mM sodium sulfite. The anode was lifted 12 cm above the outlet during electrophoresis. Other conditions are same as that in Fig. 3.

bility of the peak intensity with 1 mM Ce(IV) was not stable; the peak of PFLX could not be observed in some runs, leading to the highest coefficient of variation (CV) of 141% among all the peaks. The peak heights for 2 mM Ce(IV) showed good precision with CV of 3.75% and 2.0% for NFLX and PFLX, respectively. The standard errors of each peak were plotted as error bars in Fig. 5. Moreover, experiments showed that the peak heights with 5 mM Ce(IV) in the CL reagent decreased from run to run and no peaks could be observed after 40 min. Yellow precipitates were found at the outlet of the separation capillary and on the detection window of the reaction capillary, which might hinder the CL reaction or block the luminescence emitted. The solubility product constant of Ce(OH)4 is 2 × 10−48 [48], so Ce(IV) ion of high concentration will easily undergo hydrolysis to produce precipitates in neutral or basic aqueous solutions. In addition, the deep color of the CL solution stem from the high concentration Ce(IV) would absorb the CL emission, leading to low detection sensitivity [49].

3.4.

Effect of flow rate

The CL intensity enhanced with increasing flow rate and reached maximum at 8 ␮L min−1 ; higher flow rate led to lower detection sensitivity (Fig. 6). More CL reagent brought by the high-flow rate facilitates the CL reaction and consequently favors high-detection sensitivity. On the other hand, higher flow rate results in further dilution of the analytes and therefore the weak CL response.

3.5. Fig. 4 – Influence of buffer pH. The running buffers: 20 mM sodium citrate and 10 mM sodium sulfite titrated with citric acid to pH values of (A) 5.4; (B) 6.1; (C) 6.6. The peak * was impurities from the NFLX standard, it was approximately 3% of the NFLX standard by measuring the peak areas. Other conditions are same as that in Fig. 3.

Effect of terbium in the CL reagent

Tb(III) acted as sensitizer by forming complex with FQs; the complex accepts energy from the excited SO2 * [40,41]. Hence, the concentration of Tb(III) influences the CL intensity in the studied system. Experiments showed that both the peak heights of NFLX and PFLX showed a convex profile as a function of Tb(III) concentration, and reached maximum at 4 mM

a n a l y t i c a c h i m i c a a c t a 6 2 3 ( 2 0 0 8 ) 231–237

235

Fig. 6 – Influence of flow rate. The anode was lifted 18 cm above the outlet during electrophoresis. Other experimental conditions were same as that in Fig. 3.

(figure not shown). So Tb(III) of 4 mM was selected for further investigation.

3.6.

Method validation and application

Under the optimal conditions, the linear regression equations of peak area (y, ␮V s) versus the concentrations of NFLX and PFLX (x = 0.4–80 ␮g mL−1 , n = 13) were y = 1798(±56)x + 2581(±1848) (correlation coefficient, r2 = 0.9947) of and y = 1303(±47)x + 2115(±1536) (r2 = 0.9928), respectively. Table 1 summarized relative standard deviations (R.S.D.s) and limits of detection (LODs) for NFLX and PFLX using the proposed method. While the LOD of NFLX obtained in this study was higher than 0.0048 ␮mol L−1 of Ref. [17], it was however more sensitive than 0.11 ␮g mL−1 [50] or 0.2 ␮g mL−1 [51] by UV detection. The intra-day repeatability was calculated from consecutive five injections each day and the inter-day reproducibility was assessed from three consecutive days. The R.S.D.s of migration time for NFLX and PFLX ranged from 2.4% to 4.0% within a day (n = 5) and from 3.5% to 4.2% in 3 days (n = 15). The intra-day R.S.D.s of peak areas were all less than 14%. The maximal inter-day R.S.D. of the FQs was 18.3%, slightly higher than the recommended tolerance of 15% [52], but still less than 23% suggested by the European Union [53]. The method was also applied to analyze urine samples fortified with NFLX and PFLX at 2 ␮g mL−1 each, and a HPLC–UV analysis was carried out for comparison (Fig. 7). The analysis time of HPLC was about 21 min, and peaks corresponding to constituents in the sample matrix presented in the chromatogram of blank urine sample. In addition, the peak of NFLX partially merged with the peak * of the urine matrix, leading to errors in quantitation. However, the analytes were baseline separated within 11 min by the CE method, shorter than the HPLC method. Moreover, the electropherogram of the blank urine sample is clean, suggesting the high selectivity of the CL detector. It should be mentioned that the migration times of

Fig. 7 – Comparison of HPLC–UV and CE–CL analysis of fortified urine samples. HPLC–UV: (A) urine sample spiked with FQs of 2 ␮g mL−1 each and (B) blank urine sample. The unidentified peaks * and ** generate from the blank urine sample. The full chromatogram of trace A is put in the inset for clarity. CE–CL: (C) urine sample spiked with 2 ␮g mL−1 FQs and digitally filtered with filter width of 19; (D) the original electropherogram of trace C; (E) blank urine sample. In both the blank and the fortified samples, the urine collected from the healthy volunteer was 100-fold diluted with deionized water.

the analytes changed slightly with the fortified urine samples due to the influence of sample matrix. However, the R.S.D.s of the migration times of NFLX and PFLX (at 2 ␮g mL−1 each) for three consecutive runs were 0.77% and 0.88%, respectively, suggesting good precision. Fig. 7 also depicts that digital filtering does not result in serious deformation of the peaks, and the filtering width of 19 is suitable for the quantitative calculation. The recoveries of NFLX and PFLX spiked in urine samples were 90.6% and 93.2%, respectively, for the CE–CL; and were 99.3% and 82.0%, respectively, for HPLC–UV. Considering the residual levels of the FQs in human urine after administration [40] as well as the limit of 0.15 ␮g mL−1 suggested by Toussaint et al. [54] for FQs that are not listed in the Regulation 2377/90, we suggest that the established CE–CL method is suitable for quick determination of FQ residues in real sample.

236

a n a l y t i c a c h i m i c a a c t a 6 2 3 ( 2 0 0 8 ) 231–237

Table 1 – R.S.D.s and LODs of the determination of NFLX and PFLX R.S.D. (intra-day, %, n = 5) Peak area NFLX PFLX a

4.

12.0–13.8 10.9–13.6

Migration time 2.4–3.2 3.1–4.0

R.S.D. (inter-day, %, n = 5 × 3) Peak area 18.3 13.7

LODa (␮g mL−1 )

Migration time 3.5 4.2

0.057 0.084

LOD was defined as the S/N to be 3.

Conclusions

A CE–CL method for analyzing NFLX and PFLX was developed by employing Ce(IV) ion as the oxidant, sodium sulfite as reductant and Tb(III) as sensitizer. The high-separation efficiency of the CE technique and the high selectivity of the CL system towards the analytes enable the method to quickly assay FQ residues in real complex matrix. The method can be also applied to other types of quinolones that can form complex with Tb(III) during the CL reaction and undergo intramolecule energy transfer.

Acknowledgements This work was supported by the National Natural Science Foundation of China (20575009 and 20775008), Beijing Programme Foundation for Excellent Talents (20061D0503100307) and the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry.

references

[1] D.C. Hooper, Biochim. Biophys. Acta: Gene Struct. Expr. 1400 (1998) 45. ˜ F.J. Barragan, ´ M. Callejon, ´ F. De la Rosa, [2] J.A. Ocana, Microchim. Acta 144 (2004) 207. [3] M. Hernandez, F. Borrull, M. Calull, J. Chromatogr. B 742 (2000) 255. [4] F. Canada-Canada, A. Espinosa-Mansilla, A.M. de la Pena, J. Sep. Sci. 30 (2007) 1242. [5] D.C. Hooper, Clin. Infect. Dis. 31 (2000) S24. [6] Commission of the European Communities, Off. J. Eur. Commun. L224 (1990) 1. [7] Commission of the European Communities, Off. J. Eur. Commun. L139 (2004) 1. [8] F. Prost, W. Thormann, Electrophoresis 22 (2001) 3270. [9] J. Barbosa, D. Barron, E. Jimenez-Lozano, V. Sanz-Nebot, Anal. Chim. Acta 437 (2001) 309. [10] E. Jimenez-Lozano, D. Roy, D. Barron, J. Barbosa, Electrophoresis 25 (2004) 65. [11] D. Barron, E. Jimenez-Lozano, J. Barbosa, J. Chromatogr. A 919 (2001) 395. [12] D. Barron, E. Jimenez-Lozano, J. Barbosa, Anal. Chim. Acta 415 (2000) 83. [13] C.L. Cheng, C.H. Fu, C.H. Chou, J. Chromatogr. B 856 (2007) 381. [14] D. Barron, E. Jimenez-Lozano, S. Bailac, J. Barbosa, Anal. Chim. Acta 477 (2003) 21. [15] Y.X. Fan, X. Gan, S. Li, W.D. Qin, Electrophoresis 28 (2007) 4101. [16] M. Himmelsbach, W. Buchberger, C.W. Klampfl, Electrophoresis 27 (2006) 1220.

[17] B.Y. Deng, C.N. Su, Y.H. Kang, Anal. Bioanal. Chem. 385 (2006) 1336. [18] D. Barron, E. Jimenez-Lozano, J. Cano, J. Barbosa, J. Chromatogr. B 759 (2001) 73. [19] M. Hernandez, C. Aguilar, F. Borrull, M. Calull, J. Chromatogr. B 772 (2002) 163. [20] A.F. Faria, M.V.N. de Souza, M.V. de Almeida, M.A.L. de Oliveira, Anal. Chim. Acta 579 (2006) 185. [21] L. Wang, X.P. Wu, Z.H. Xie, J. Sep. Sci. 28 (2005) 1143. ¨ [22] C. Horstkotter, E. Jimenez-Lozano, D. Barron, J. Barbosa, G. Blaschke, Electrophoresis 23 (2002) 3078. [23] F.J. Lara, A.M. Garcia-Campana, F. Ales-Barrero, J.M. Bosque-Sendra, L.E. Garcia-Ayuso, Anal. Chem. 78 (2006) 7665. [24] A. Juan-Garcia, G. Font, Y. Pico, Electrophoresis 27 (2006) 2240. [25] A. Garces, A. Zerzanova, R. Kucera, D. Barron, J. Barbosa, J. Chromatogr. A 1137 (2006) 22. [26] M. Elbanowski, B. Makowska, K. Staninski, M. Kaczmarek, J. Photochem. Photobiol. A: Chem. 130 (2000) 75. [27] X. Wang, H.C. Zhao, L.H. Nie, L.P. Jin, Z.L. Zhang, Anal. Chim. Acta 445 (2001) 169. [28] S.L. Chen, Y. Liu, H.C. Zhao, L.P. Jin, Z.L. Zhang, Y.Z. Zheng, Luminescence 21 (2006) 20. [29] R.W. Abbott, A. Townshend, R. Gill, Analyst 111 (1986) 635. [30] L. Gamiz-Gracia, A.M. Garcia-Campana, J.J. Soto-Chinchilla, J.F. Huertas-Perez, A. Gonzalez-Casado, Trend Anal. Chem. 24 (2005) 927. [31] P.J. Fletcher, K.N. Andrew, S. Forbes, P.J. Worsfold, Anal. Chem. 75 (2003) 2618. [32] D.W. King, W.J. Cooper, S.A. Rusak, B.M. Peake, J.J. Kiddle, D.W. O’Sullivan, M.L. Melamed, C.R. Morgan, S.M. Theberge, Anal. Chem. 79 (2007) 4169. [33] A.M. Garcia-Campana, L. Gamiz-Gracia, W.R.G. Baeyens, F.A. Barrero, J. Chromatogr. B 793 (2003) 49. [34] F.J. Lara, A.M. Garcia-Campana, L. Gamiz-Gracia, J.M. Bosque-Sendra, F. Ales-Barrero, Electrophoresis 27 (2006) 2348. [35] J.C. Dumke, M.A. Nussbaum, Anal. Chem. 79 (2007) 1262. [36] C.F. Peng, T.M. Huo, L.Q. Liu, X.G. Chu, C.L. Xu, Electrophoresis 28 (2007) 970. [37] S.L. Chen, H.C. Zhao, C.Y. Sun, N. Lian, L.P. Jin, Anal. Lett. 35 (2002) 1705. [38] L. Ning, H.C. Zhao, C.Y. Sun, S.L. Chen, Y. Lu, L.P. Jin, Microchem. J. 74 (2003) 223. [39] J. Stauff, W. Jaeschke, Z. Naturforsch. B 33 (1978) 293. [40] L. Yi, H.C. Zhao, S. Chen, L.P. Jin, D.D. Zheng, Z.L. Wu, Talanta 61 (2003) 403. [41] L.-H. Nie, H.-C. Zhao, X. Wang, L. Yi, Y. Lu, L.-P. Jin, H.-M. Ma, Anal. Bioanal. Chem. 374 (2002) 1187. [42] S. Cristoni, M. Cantu, L.R. Bernardi, P. Gerthoux, P. Mocarelli, M. Brambilla, E. Gonella, F. Guidugli, J. Mass Spectrom. 40 (2005) 1609. [43] H. Ziegler, Appl. Spectrosc. 35 (1981) 88. [44] I. Kovacevic, M. Pokrajac, B. Miljkovic, D. Jovanovic, M. Prostran, J. Chromatogr. B 830 (2006) 372. [45] C.-E. Lin, Y. Deng Jr., W.-S. Liao, S.-W. Sun, W.-Y. Lin, C.-C. Chen, J. Chromatogr. A 1051 (2004) 283.

a n a l y t i c a c h i m i c a a c t a 6 2 3 ( 2 0 0 8 ) 231–237

[46] http://www.wipo.int/pctdb/en/wo.jsp?IA=WO2000001365 &DISPLAY=DESC. [47] Calculated using Advanced Chemistry Development (ACD/Labs) Software V8.14 for Solaris, ACD/Labs, 1994–2008. [48] J.A. Dean, Lange’s Handbook of Chemistry, McGraw-Hill, Inc., New York, 1999. [49] X.L. Cheng, L.X. Zhao, X. Wang, J.M. Lin, Anal. Sci. 23 (2007) 1189. [50] A. Alnajjar, A.M. Idris, H.H. AbuSeada, Microchem. J. 87 (2007) 35.

237

[51] H.X. Lu, X.P. Wu, Z.H. Xie, X.C. Lin, L.Q. Guo, C. Yan, G.N. Chen, J. Sep. Sci. 28 (2005) 2210. [52] US Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research, Center for Veterinary Medicine; Guidance for Industry, Bioanalytical Method Validation, May 2001, http://www.fda.gov/cder/guidance/index.html. [53] Commission of the European Communities, Off. J. Eur. Commun. L118 (1993) 64. [54] B. Toussaint, M. Chedin, G. Bordin, A.R. Rodriguez, J. Chromatogr. A 1088 (2005) 32.