Direct Extraction of Penicillin G and Derivatives from Aqueous

The samples were filtered through a 0.45 μm filter (Whatman, Maidstone, U.K.) to ...... López-Rivadulla , José M. López-Vilariño , M. Victoria Go...
0 downloads 0 Views 318KB Size
Download PDF
Anal. Chem. 2007, 79, 695-701

Direct Extraction of Penicillin G and Derivatives from Aqueous Samples Using a Stoichiometrically Imprinted Polymer Javier L. Urraca,† Marı´a C. Moreno-Bondi,*,† Andrew J. Hall,‡ and Bo 1 rje Sellergren*,§

Department of Analytical Chemistry, Faculty of Chemistry, Universidad Complutense de Madrid, E-28040 Madrid, Spain, Institute of Pharmacy, Chemistry and Biomedical Sciences, University of Sunderland, Wharncliffe Street, Sunderland SR1 3SD, United Kingdom, and Institut fu¨r Umweltforschung, Universita¨t Dortmund, Otto Hahn Strasse 6, D-44221 Dortmund, Germany

A molecularly imprinted polymer (MIP) prepared using penicillin G procaine salt as the template (PENGp) and a stoichiometric quantity of urea-based functional monomer to target the single oxyanionic species in the template molecule has been applied to the development of a molecularly imprinted solid-phase extraction (MISPE) procedure for the selective preconcentration of β-lactam antibiotics (BLAs) from environmental water samples. Various parameters affecting the extraction efficiency of the polymer have been evaluated to achieve the selective preconcentration of the antibiotics from aqueous samples and to reduce nonspecific interactions. This resulted in an MISPE-HPLC method allowing the direct extraction of the analytes from the sample matrix with a selective wash using just 10% (v/v) organic solvent. On the basis of UV detection only, the method showed good recoveries and precision, ranging between 93% and 100% (RSD 3.8-8.9%, n ) 3) for tap water and between 90% and 100% (RSD 4.2-9.1%, n ) 3) for river water fortified with 30 or 60 µg L-1 (50 mL samples) penicillin G, penicillin V, nafcillin, oxacillin, cloxacillin, and dicloxacillin, suggesting that this MIP can be successfully applied to the direct preconcentration of BLAs in environmental water samples.

INTRODUCTION Water quality control is one of the most relevant issues in environmental analytical chemistry today. Antibiotics can be released into the environment after their application in human and veterinary medicine or their use as growth promoters in animal husbandry, fish farming, and other fields. In recent years, antibiotics and bacteria resistant to them have been detected in the effluent of pharmaceutical companies and hospitals, in municipal wastewater, and in surface water or in groundwater, in both Europe and North America, at concentrations ranging from * To whom correspondence should be addressed. E-mail: mcmbondi@ quim.ucm.es (M.C.M.-B.); [email protected] (B.S.). † Universidad Complutense de Madrid. ‡ University of Sunderland. § Universita¨t Dortmund. 10.1021/ac061622r CCC: $37.00 Published on Web 12/01/2006

© 2007 American Chemical Society

ng L-1 to µg L-1.1-5 Incomplete removal of these drugs in municipal sewage treatment plants (STPs) is considered to be one of the main sources of antibiotic entrance into the environment.1 One of the main interests in the determination of antibiotic residues in the environment arises from the fact that they are suspected of being responsible for the appearance of growing resistance of pathogenic bacteria, from their ecotoxic effects, and because they can enter the food chain via uptake by plants or translocation into groundwater.1,6 The presence of resistant bacteria seems to be an important source of resistance in the environment, although its possible impact is still under investigation. The presence of antibiotic residues in live animals and animal products is legislated in Europe (Council Directive 96/23/EC), and countries must monitor levels for safety assessment. However, these antimicrobials have not been included in the list of priority and hazardous substances in the Water Framework Directive of the European Union.7,8 So, there is an increasing demand for sensitive and reliable analytical methods for the analysis of antibiotics in groundwater and surface water samples. Penicillins are a group of β-lactam antibiotics (BLAs) that have been widely applied in human and veterinary medicine since their introduction in 1941. These antibiotics can be classified into three groups: (a) natural penicillins, including penicillin G (PENG) and penicillin V (PENV), (b) penicillinase-resistant penicillins, including methicillin (METH), nafcillin (NAFCI), oxacillin (OXA), cloxacillin (CLOX), dicloxacillin (DICLOX), and mecillinam (MECI), and (c) broad-spectrum penicillins, including ampicillin (AMPI), amoxicillin (AMOX), and hetacillin (HETA).9 (1) Junker, T.; Alexy, R.; Knacker, T.; Kummerer, K. Environ. Sci. Technol. 2006, 40, 318-324. (2) Kummerer, K. J. Antimicrob. Chemother. 2004, 54, 311. (3) Giger, W.; Alder, A. C.; Golet, E.; Kohler, H. P. E.; McArdell, C. S.; Molnar, E.; Siegrist, H.; Suter, M. J. F. Chimia 2003, 57, 485. (4) Kolpin, D.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Environ. Sci. Technol. 2002, 36, 1202. (5) Golet, E. M.; Strehler, A.; Alder, A. C.; Giger, W. Anal. Chem. 2002, 74, 5455. (6) Hirsch, R.; Ternes, T. A.; Haberer, K.; Mehlich, A. J. Chromatogr., A 1998, 815, 213. (7) Water Framework Directive (2000/60/EC); European Union: Brussels, Belgium, 2000.

Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 695

The literature available on the analysis of penicillins in environmental samples is very limited compared to that of other antibiotics, such as sulfonamides, tetracyclines, fluoroquinolones, or macrolides.10,11-13 The analysis of BLAs in environmental samples is usually accomplished using liquid chromatography (LC) with mass spectrometric (MS) detection.10 Several methods have been described, for routine analysis, using solid-phase extraction (SPE) and LC with UV detection.14 However, there is a lack of SPE sorbents able to selectively preconcentrate β-lactam antibiotics with good recoveries from complex environmental waters.14 An inmunosensor based on a competitive fluorescent immunoassay has also been described.15 The application of molecular imprinted polymers (MIPs) to SPE, namely, molecularly imprinted solid-phase extraction (MISPE), is an interesting alternative to achieve selective extraction of the target compound when the commonly applied sorbents lack selectivity.16 One of the main limitations associated with the broad applicability of molecularly imprinted polymers in SPE is their poor performance in aqueous media. The competition between the analyte and the solvent for the MIP functional groups as well as the differences in polymer swelling experienced in the porogen solvent and in water may account for this effect. In most MISPE procedures, when the sample is aqueous, the polymer will retain the analyte during the loading step by selective imprinting binding and nonspecific adsorption. To achieve a selective extraction, a cleanup step must be introduced using an organic solvent, prior to the elution step, in which the imprint binds to the analyte selectively and the nonspecific interactions with the polymer are minimized or not present. Indeed, this step is much more critical in MISPE methods than in conventional SPE. Alternatively, a precolumn can be used to capture the analyte from the aqueous sample, and afterward it can be transferred to an organic solvent in which the MIP-analyte interaction is selective. Obviously, any of these approaches increments the complexity, price, and environmental friendliness, when organic solvents are required, of commercial SPE procedures, and great effort has been focused on the development of aqueous compatible MIPs capable of direct analyte preconcentration. We have recently described17 the synthesis of an imprinted polymeric receptor for the class-selective recognition of BLAs, from the penicillin family, able to effectively retain such antimicrobials in aqueous samples. The polymer was prepared using a “stoichiometric” noncovalent imprinting approach18 with penicillin (8) Cahill, J. D.; Furlong, E. T.; Burkhardt, M. R.; Kolpin, D.; Anderson, L. G. J. Chromatogr., A 2004, 1041, 171. (9) Golet, E. M.; Alder, A. C.; Hartmann, A.; Ternes, T. A.; Giger, W. Anal. Chem. 2001, 73, 3632. (10) Golet, E. M.; Strehler, A.; Alder, A. C.; Giger, W. Anal. Chem. 2002, 74, 5455. (11) Hartig, C.; Storm, T.; Jekel, M. J. Chromatogr., A 1999, 854, 163. (12) http://europa.eu.int/comm/environment/water (accessed Nov 22, 2004). (13) Botsoglou, N. A., Fletouris D. J., Eds. Drugs Residues in Foods: Pharmacology, Food Safety, and Analysis; Marcel Dekker, Inc.: New York, 2000; p 41. (14) Benito-Pen ˜a, E.; Partal-Rodera, A. I.; Leo´n-Gonza´lez, M. E.; Moreno-Bondi. M. C. Anal. Chim. Acta 2006, 556, 422. (15) Benito-Pen ˜a, E.; Moreno-Bondi, M. C.; Aparicio, S.; Orellana, G.; Cederfur, J.; Kempe, M. Anal. Chem. 2006, 78, 2019. (16) Alexander, C.; Anderson, H. S.; Anderson, L. I.; Ansell, R. J.; Kirsch, N.: Nicholls, I. A.; Whitcombe, M. J. J. Mol. Recognit. 2006, 19, 106. (17) Urraca, J. L. Hall, A. J. Moreno-Bondi M. C.; Sellergren, B. Angew. Chem., Int. Ed. 2006, 45, 5158. (18) Hall, A. J.; Manesiotis, P.; Emgenbroich, M.; Quaglia, M.; De, Lorenzi, E.; Sellergren, B. J. Org. Chem. 2005, 70, 1732.

696

Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

G procaine salt as the template, 1-(4-vinylphenyl)-3-(3,5-bis(trifluromethyl)phenyl)urea and methacrylamide as functional monomers, ethylene glycol dimethacrylate (EDMA) as the crosslinker (ratio 0.5:0.5:1:20), and acetonitrile as the porogen. The retention mechanism of this material has been attributed to the strong and stoichiometric interactions between the deprotonated carboxylic acid groups of the analytes and the urea-based monomer,17 and its application may allow one of the main limitations associated with the application of the BLA-selective MIPs described in the literature to real sample analysis to be overcome, i.e., the requirement of a direct recognition effect in aqueous media.19,20 In the present paper we describe for the first time the application of an MIP imprinted with PENG to the direct extraction of low levels of BLAs (AMOX, AMPI, PENG, PENV, CLOX, DICLO, NAFCI, OXA) (Figure 1) from river and tap water samples. Penicillins were selectively retained in the polymer in the loading step without the need of further cleanup using a pure organic solvent. The method has been validated using spiked river water samples using LC-UV detection only. EXPERIMENTAL SECTION 1. Chemicals. The urea-based functional monomer 1-(4vinylphenyl)-3-(3,5-bis(trifluromethyl)phenyl)urea (1) was prepared as described previously.18 EDMA was purchased from Aldrich and purified prior to use as follows: EDMA was washed sequentially with 10% NaOH (aqueous), water, and then brine. After being dried over MgSO4, it was distilled under reduced pressure to give an inhibitor-free monomer. Methacrylamide was purchased from Sigma-Aldrich (St. Louis, MO). ABDV was obtained from Wako (Neuss, Germany) and used as received. The antiobiotics penicillin G potassium salt (PENG) and procaine salt (PENGp), penicillin V potassium salt (PENV), AMOX, nafcillin sodium salt (NAFCI), cloxacillin sodium salt (CLOX), dicloxacillin sodium salt (DICLOX), OXA, and AMPI were supplied by Bayer AG (Leverkusen, Germany) and used as received. Trifluoroacetic acid (TFA) (HPLC grade, 99%) and humic acids were from Fluka (Buchs, Switzerland). Tetra-n-butylammonium hydrogen sulfate (TBA) (98%) and sodium chloride were from Merck (Darmstadt, Germany). HPLC grade acetonitrile and methanol were purchased from SDS (Peypin, France), and HPLC water was purified using a Milli-Q system (Millipore, Bedford, MA). All solutions prepared for HPLC were passed through a 0.45 µm nylon filter before use. HEPES was supplied by Aldrich (Steinheim, Germany), and a buffer solution, pH 7.5, was prepared by dissolving 23.83 g in 1 L of Milli-Q water (0.1 M). 2. Apparatus. The pH of the buffer solutions and samples was adjusted with an Orion 710A pH/ISE meter (Beverly, MA). The chromatographic system consisted of an HP-1100 series highperformance liquid chromatograph from Agilent Technologies (Palo Alto, CA) equipped with a quaternary pump, an on-line degasser, an autosampler, an automatic injector, a column thermostat, and a diode-array (DAD) detector. A peristaltic pump miniplus 3 (Gilson) was applied for sample preconcentration in the cartridges. (19) Skudar, K.; Bru ¨ ggemann, O.; Wittelsberger, A.; Ramstro ¨m, O. Anal. Commun. 1999, 36, 327. (20) Cederfur, J.; Pei, Y.; Zihui, M.; Kempe, M. J. Comb. Chem. 2003, 5, 67.

Figure 1. Chemical structures, acronyms, and pKa values of the investigated BLAs.

Chromatographic separation of the BLAs was performed on a LUNA C18 (2) (150 × 4.6 mm, 5 µm) HPLC column protected by an RP18 guard column (4.0 × 3.0 mm, 5 µm), both from Phenomenex (Torrance, CA). A gradient program was used with the mobile phase, combining solvent A (Milli-Q water with 0.01% TFA) and solvent B (ACN with 0.01% TFA) as follows: 0% B (3 min), 0-37% B (5 min), 37% B (11 min), 37-67% B (5 min), 67% B (5 min). Analyses were performed at a flow rate of 1.5 mL min-1, and the column temperature was kept at 35 °C. The injection volume was 200 µL, and all the compounds eluted within 24 min. The UV detector wavelength was set at 220 nm. Quantification was performed using external calibration and peak area measurements. Linear calibration graphs were obtained in the 75 µg L-1 to 5 mg L-1 range for all the antibiotics (R2 > 0.999). 3. Polymer Preparation. The polymers were prepared as described previously.17 Briefly, the template PENGp (286 mg, 0.5 mmol), functional monomer 1 (186 mg, 0.5 mmol), methacrylamide (84 mg, 1 mmol), EDMA (3.8 mL, 20 mmol), and the free radical initiator ABDV (44 mg, 1% w/w total monomers) were dissolved in MeCN (5.6 mL). After dissolution, the solution was transferred to a glass tube, cooled to 0 °C, and then purged with N2 for 10 min. After purging, the glass tube was sealed and polymerization initiated thermally by placing the tube in a water bath set at 40 °C. Polymerization was allowed to proceed at this temperature for 48 h. The tube was then broken and the MIP monolith removed and broken into smaller fragments. The template molecule was removed through the following sequential washing steps: MeOH (100 mL), MeOH/0.1 M HCl(aq) (9:1, v/v, 100 mL), and finally MeOH (100 mL). The wash solutions were combined and evaporated to dryness under reduced pressure. The solid residue was then weighed and examined by 1H NMR spectroscopy, showing that template removal was nearly quantitative (extract residue mass 300 mg). Thereafter, the MIP was crushed and sieved, and particles in the size range 25-50 µm

were collected for use in the SPE experiments. Prior to use, they were sedimented using MeOH/water (80:20, v/v) to remove fine particles. A control, nonimprinted polymer (NIP) was prepared in the same manner, but with omission of the template molecule. 4. Optimized Extraction Procedure of BLAs in the MIP Cartridge. Solid-phase extraction cartridges (Varian, Spain), with a 3 mL volume, were packed with 150 mg of the PENG-imprinted or the corresponding nonimprinted polymers. The cartridges were equilibrated with 10 mL of buffer (HEPES, 0.1 M, pH 7.5), and the antibiotic sample, dissolved in buffer (HEPES, 0.1 M, pH 7.5), was percolated at a constant flow rate of 0.75 mL min-1 with the aid of a peristaltic pump. The cartridges were washed with 5 mL of a mixture of buffer (HEPES, 0.1 M, pH 7.5)/AcN (90:10, v/v) to elute the nonspecifically retained compounds. Finally the antibiotics were eluted with 1 mL of a solution of 0.05 M TBA in MeOH. The cartridges were reequilibrated with 10 mL of buffer (HEPES, 0.1 M, pH 7.5) before a new application. The eluates from the MISPE column (80 µL) were mixed with 120 µL of buffer (HEPES, 0.1 M, pH 7.5) and injected into the HPLC system for analysis. 5. Humic Acid Sample Analysis. A stock solution of river humic acids (HAs) was prepared by dissolving 20 mg in aqueous NaOH (25 mM) to a final concentration of 200 mg L-1. HA solutions with concentrations ranging between 0.01 and 10 mg L-1 were made up by dilution of the stock solution, and the pH was adjusted to 7.5 with HEPES buffer. 6. River and Tap Water Sample Analysis. River and tap water samples were collected in 2.5 mL amber glass bottles prerinsed with ultrapure water and stored at 4 °C until measurement. River water samples were collected from the Guadarrama River (Madrid, Spain). The samples were filtered through a 0.45 µm filter (Whatman, Maidstone, U.K.) to remove suspended matter, and the pH was adjusted to 7.5 using HEPES and NaOH to a final buffer concentration of 0.1 M. The samples were fortified Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

697

Figure 2. Recoveries obtained for different flow rates using a 10 mL solution of PENG (0.3 mg L-1) as the percolation volume. PENG elution from the MIP cartridge was carried out with 1 mL of methanol (TBA, 0.05 M).

with the target analytes at two concentration levels, 30 and 60 µg L-1. Thereafter, the samples were passed through the MIP cartridges as described in subsection 4 of the Experimental Section. For quantification purposes nonfortified water samples were preconcentrated in the MIP cartridges, checked for not containing the analytes at the method detection limits, and spiked with the antibiotic stock solutions. The method detection limit was determined as the minimum detectable amount of an analyte in the spiked water sample giving a signal-to-noise ratio of 3. All analyses were carried out in triplicate. RESULTS AND DISCUSSION 1. Optimization of the BLA SPE Procedure. The factors evaluated to establish the optimum conditions for the SPE procedure include the study of the composition and volume of the eluting solvent, the flow rate of the loading solution, the composition of the washing solvent, and the breakthrough volume (the maximum volume that can be preconcentrated with quantitative recovery of analyte). a. Elution Solvent Optimization. The elution solvent was optimized in the first instance. To this end, 10 mL samples containing 3 µg of PENG dissolved in HEPES, 0.1 M (pH 7.5), were loaded onto the cartridge and eluted with 3 mL of MeOH (1 + 1 + 1 mL). The concentration of the antibiotic was determined for each 1 mL fraction. The use of 3 mL of MeOH allowed the quantitative recovery of PENG (101%, RSD 5%, n ) 3). To reduce the eluting volume and increase the sensitivity of the assay, a solution containing 2% formic acid in MeOH was used instead. The addition of formic acid disrupts the interactions between the polymers and the template, upon protonation of the carboxylic acid of the antibiotic, but unfortunately, the presence of this acid also favors the decomposition of PENG, leading to a decrease in the recovery yields, as shown in the Supporting Information. Due to the nature of the sorbent, anions such as hydroxide, borate, bicarbonate, etc. can be used to facilitate the elution of the penicillins and avoid the undesirable hydrolysis on the β-lactam ring observed in acidic media. We have shown previously14 that 698 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

Figure 3. Extraction recoveries (%) obtained on the MIP and the NIP for eight penicillins after the percolation of 50 mL of buffer (HEPES, 0.1 M, pH 7.5) spiked with 3 µg of each compound using a washing step with (a, top) 2.5 mL and (b, bottom) 5 mL of AcN/buffer (HEPES, 0.1 M) (10:90, v/v).

Figure 4. Curves of capacity obtained after the loading of (1) 10 mL, (2) 50 mL, and (3) 100 mL of buffer (HEPES, 0.1 M, pH 7.5) spiked with increasing amounts of PENG on the MIP (150 mg). The dotted line corresponds to a slope of 1, meaning an extraction recovery of 100%.

an ion pair forming agent, TBA, can be used to increase the elutropic strength of the MeOH eluting solutions when penicillins are preconcentrated on mixed-mode polymeric SPE sorbents thanks to formation of ion pairs with the analyte. The application of 1 mL of a 0.05 M TBA solution in MeOH allowed the quantitative recovery (105%, RSD 6%) of the antibiotic (check the

Figure 5. Extraction recoveries (%) obtained on the MIP for penicillins after the percolation of 10 mL of buffer (HEPES, 0.1 M, pH 7.5) spiked with 3 µg of each antibiotic, in the presence of different concentrations of HAs, using a washing step with 5 mL of AcN/buffer (HEPES, 0.1 M) (10:90, v/v). Table 1. Limits of Detection (LODs) for the Eight BLAs after the Percolation of 50 mL of Water Calculated as 3 Times the Signal-to-Noise Ratio LOD (µg L-1) anitibiotic

Milli Q water

tap water

river water

AMOX AMPI PENG PENV OXA CLOX NAFCI DICLOX

0.89 0.65 0.75 0.98 0.38 0.40 0.52 0.63

2.8 2.9 1.7 2.4 0.9 1.1 1.0 1.2

5.8 5.3 1.3 2.6 1.9 1.8 1.7 2.8

Supporting Information), so it was selected for further experiments. The direct injection of the SPE extract in the HPLC column gave rise to a significant peak distortion in the chromatographic separation, even if smaller sample volumes were injected into the column. This behavior was attributed to the higher elution strength of the eluate (0.05 M TBA in methanol) with respect to the mobile phase. One way to avoid this problem would be to evaporate the organic solvent and reconstitute the extract in the mobile phase.14,21 However, penicillins are very unstable under these conditions, and a dilution of the eluate, 80 µL, with 120 µL of Milli-Q water was selected as the noneluting focusing solvent. b. Effect of the Flow Rate of the Loading Solution. Figure 2 shows the effect of the flow rate of the loading solution on the recovery of PENG when 10 mL of a solution of the antibiotic (0.3 mg L-1, in HEPES, 0.1 M, pH 7.5) was loaded into the MIP cartridge. An average extraction recovery of 102% (RSD 3%, n ) 3) was obtained at flow rates e0.75 mL min-1. However, the use of higher rates led to a continuous decrease in the recovery values as the interaction time between the analytes and the sorbent was decreased. A loading rate of 0.75 mL min-1 was selected for further experiments. c. Washing Solvent Selection. Penicillins can be present in aqueous solutions as neutral, anionic, or, in the case of AMOX and AMPI, intermediate forms (zwitterions), due to the presence (21) Prasad, B. B.; Banerjee, S. React. Funct. Polym. 2003, 55, 159.

Figure 6. Chromatogram obtained before (---) and after (s) percolation of 50 mL of (a, top) tap water and (b, bottom) river water, spiked with the eight antibiotics (60 µg L-1): (1) amoxicillin; (2) ampicillin; (3) penicillin G; (4) penicillin V; (5) oxacillin; (6) cloxacillin; (7) nafcillin; (8) dicloxacillin.

of carboxylic and amino groups in the molecules (Figure 1). Therefore, their extraction behavior will be pH dependent. Regarding their stability, β-lactams are readily degraded under strongly acidic and basic conditions as a result of the hydrolysis of the nucleophilic β-lactam ring.22 As described in our previous work,17 the nonspecific interaction between the analytes and the imprinted polymer can be minimized in the presence of a mixture of buffer (HEPES, 0.1 M, pH 7.5)/ AcN (90:10, v/v). Thus, this solvent was tested for the washing step. To determine the optimum washing volume, a 50 mL water sample (HEPES, 0.1 M, pH 7.5), spiked with 3 µg of AMOX, AMPI, PENG, PENV, NAFCI, OXA, CLOXA, and DICLOXA, was percolated through the MIP and the NIP cartridges, and different volumes (2.5 or 5.0 mL) of a mixture of buffer (HEPES, 0.1 M, pH 7.5)/AcN (90:10, v/v) were applied in the washing step. Thereafter, 1 mL of 0.05 M TBA in methanol was applied for elution, and the eluates were analyzed by HPLC. The results are collected in Figure 3. The use of 2.5 mL of washing solvent allowed recoveries higher than 90% (RSD 4.5-6.3%, n ) 3) in the MIP cartridge for all the (22) Reverte´, S.; Borrull, F.; Pocurull, E.; Marce´, R. M. J. Chromatogr., A 2003, 1010, 225.

Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

699

Table 2. Average Recoveries (R, %) and Relative Standard Deviations (RSDs, %, n ) 3) Obtained after Solid-Phase Extraction of 50 mL of Tap and River Water Samples Spiked at 30 and 60 µg L-1 Concentration Levels tap water

river water

MISPE-LC-UV

antibiotic AMOX AMPI PENG PENV OXA CLOX NAFCI DICLOX

MISPE-LC-UV

spiked level (µg L-1)

level found (µg L-1), mean (RSD, %)

R (%)

30 60 30 60 30 60 30 60 30 60 30 60 30 60 30 60

14.3 (7.8) 30.6 (8.8) 21.3 (8.4) 44.5 (9.7) 28.9 (7.6) 57.0 (8.4) 29.9 (8.0) 60.1 (4.8) 28.8 (3.8) 58.1 (8.9) 29.4 (5.8) 59.2 (5.9) 28.4 (5.6) 57.4 (7.0) 28.0 (6.4) 55.7 (7.9)

48 51 71 74 96 95 97 100 96 97 98 99 95 96 93 93

spiked level (µg L-1)

level found (µg L-1), mean (RSD, %)

R (%)

30 60 30 60 30 60 30 60 30 60 30 60 30 60 30 60

11.3 (15.0) 24.7 (11.7) 23.7 (7.5) 44.2 (7.2) 29.8 (8.4) 59.8 (7.0) 29.1 (7.6) 59.2 (4.4) 28.6 (4.2) 57.9 (9.1) 28.9 (4.5) 59.1 (6.6) 28.1 (6.0) 57.9 (7.2) 27.0 (5.2) 56.2 (7.6)

38 41 79 74 99 100 97 99 95 96 96 98 94 96 90 94

BLAs tested, except AMOX and AMPI. However, the use of the NIP sorbent rendered recoveries in the range 79-86% (RSD 1.96.6%, n ) 3) for PENV NAFCI, OXA, CLOXA, and DICLOXA and much lower recoveries in the case of AMOX (0%), AMPI (32%), and PENG (18%). When a washing volume of 5 mL was applied instead, the nonspecific interactions were significantly reduced for all the BLAs. In this case, the extraction recoveries in the NIP ranged between 0 and 28% (RSD 2.9-7.3%, n ) 3) for all the antibiotics, whereas retention on the MIP was excellent for PENG, PENV, OXA, CLOX, NAFCI, and DICLOX (recoveries between 92 and 99%, RSD 3.0-5.1%, n ) 3) and decreased again for AMPI (72%) and AMOX (50%). In conclusion, a volume of 5 mL of buffer (HEPES, 0.1 M, pH 7.5)/AcN (90:10, v/v) was selected for the washing step. d. Evaluation of the PENG-MIP Capacity. The characterization of an MISPE procedure requires the evaluation of the cartridge capacity, i.e., the maximum amount of a compound that can be retained by the sorbent. The determination of this parameter was carried out by percolating 10, 50, and 100 mL of buffer (HEPES, 0.1 M, pH 7.5) spiked with increasing amounts PENG on the MIP cartridge. Figure 4 shows the amount of penicillin bound on the MIP as a function of the amount loaded into the cartridge. The curve obtained with a percolated volume of 10 mL presents a linear part for the lowest percolated amounts of PENG (0-250 µg) corresponding to an extraction recovery of 100%. As indicated by Chapuis et al.23 for a reliable quantitative method, the amount of penicillins in the samples should not be over 250 µg or 1667 µg g-1 polymer. The percolation of antibiotic amounts higher than 250 µg led to a decrease in the recovery yields, and the amount retained by the MIP tended toward a plateau, indicating a saturation of the binding cavities, at 310 µg of PENG. As the percolated volume was increased to 50 and 100 mL the curves tended to saturate at lower amounts of introduced

antibiotic. Recoveries on the order of 100% were obtained, with percolated volumes of 50 and 100 mL, for PENG introduced amounts lower that 40 and 10 µg, respectively. A plateau was obtained at PENG amounts of 290 and 210 µg for 50 and 100 mL percolated volumes, respectively. This behavior has been attributed23 to the presence of binding sites of different energies in the polymer. Thus, each of the curves represented in Figure 4 may be considered as the sum of the capacity curves corresponding to the binding of PENG to binding sites of different energies. In this way, as the volume of the loading solution increases, the breakthough volume for some binding sites is attained and the number of accessible binding sites is diminished, leading to a decrease in the polymer capacity. 2. Influence of the Presence of Humic Acids in the Sample. Humic substances are ubiquitous in the environment and are formed during the microbiological and abiotic transformations of animal and plant materials. Among them, HAs represent the fraction of humic substances composed of long-chain molecules, of high molecular weight (average Mr ) 2000-3000), that are dark brown in color and are insoluble in water under acidic conditions (pH < 2), but are soluble at higher pH. Their presence may affect a variety of chemical, physical, and biological reactions in the aquatic environment.24 Recently, several research trials conducted worldwide have shown positive results concerning the use of these substances as organic additives in animal feed in place of antibiotics and hormones. They can bind to both positively and negatively charged ions and behave as powerful free radical scavengers and natural antioxidants. For the application of the MIP cartridge to the preconcentration of BLAs in environmental waters it was important to evaluate the effect of dissolved HAs on the preconcentration efficiency of the BLAs on the MIP cartridge. Figure 5 shows the recovery rates of eight BLAs after MISPE of 10 mL aqueous samples (HEPES, 0.1 M, pH 7.5) spiked with

(23) Morin, R. B., Gorman, M., Eds. Penicillins and Cephalosporins; Academic Press: New York, 1982; Vol. 1.

(24) Chapuis, F.; Pichon, V.; Lanza, F.; Sellergren, B.; Hennion, M.-C. J. Chromatogr., B 2004, 804, 93.

700 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

3 µg of each antibiotic and increasing concentrations of HAs (0.01-10 mg L-1).25 The presence of up to 0.01 mg L-1 HAs did not significatively affect the retention of PENG, PENV, NAFCI, OXA, CLOXA, and DICLOXA. Recoveries over 79% (RSD 4.4-6.2%) were obtained even in the presence of 1 mg L-1 HAs. However, the retention of the more polar antimicrobials, AMPI and, especially AMOX, decreased considerably in the presence of these compounds. Indeed, upon addition of 10 mg L-1 HAs, the recoveries decreased dramatically and new peaks were observed in the chromatogram. It has been described in the literature that humic acids can act as singlet oxygen photosensitizers, enhancing the photodegradation rate of amoxicillin.26 This effect could explain the appearance of new peaks in the chromatograms corresponding to the eluates of the solutions spiked with HA concentrations higher than 0.1 mg L-1, which might correspond to the photodegradation products. The singlet molecular oxygen produced by humic acids is probably responsible for the photodegradation of AMOX and AMPI as these two antibiotics bear singlet oxygenreactive groups (phenol and amine, respectively).27 However, for the other antibiotics, this effect does not seem to play an important role in the recovery rate. Thus, for PENG, PENV, NAFCI, OXA, CLOXA, and DICLOXA, recoveries are on the order of 70% in the presence of 10 mg L-1 HAs. This decrease may be explained considering the interaction of the humic acids with the carboxylate groups of the antibiotic, thus preventing their retention on the polymer, as has been observed for other xenobiotic compounds.28 3. Analysis of River and Tap Waters. The presence of matrix components in real samples may significantly decrease the preconcentration efficiency of antibiotics on SPE sorbents.14 To evaluate the applicability of the optimized MISPE procedure to real sample water analysis, tap and river water samples (50 mL) were spiked with 1.5 and 3 µg of each BLA and analyzed using the optimized methodology. The detection limits obtained after the percolation of 50 mL samples are collected in Table 1. Figure 6 shows the chromatograms obtained using the MISPE procedure (50 mL) on spiked (60 µg L-1) tap and river water as well as those of the corresponding nonpreconcentrated samples. (25) Aiken, G. R.; McKnight, D. M.; Wershaw, R. L.; MacCarthy, P. Geochemistry and Characterization; Wiley: NewYork, 1985. (26) Nagao, S.; Matsunaga, T.; Suzuki, Y.; Ueno, T.; Amano, H. Water Res. 2003, 37, 4159. (27) Andreozzi, R.; Caprio, V.; Ciniglia, C.; De Campdore´, M.; Lo Giudice, R.; Marotta R.; Zuccato, E. Envirom. Sci. Technol. 2004, 38, 6832. (28) DeRosa, M. C.; Crutchley, R. J. Coord. Chem. Rev. 2002, 233, 351. (29) Thiele-Bruhn, S.; Seibicke, T.; Shulten, H.-R.; Leinweber, P. J. Environ. Qual. 2004, 33, 1331. (30) Ahrer, W.; Scherwenk, E.; Buchberger, W. J. Chromatogr., A 2001, 910, 69. (31) Sacher, F.; Lange, F. T.; Brauch, H.-J.; Blankenhorn, I. J. Chromatogr., A 2001, 938, 199. (32) Bruno, F.; Curini, R.; di-Corcia, A.; Nazzari, M.; Samperi, R. Rapid Commun. Mass Spectrom. 2001, 15, 1391.

The recoveries obtained for PENG, PENV, NAFCI, OXA, CLOXA, and DICLOXA ranged between 93% and 100% (RSD 3.88.9%, n ) 3) for tap water and between 90% and 100% (RSD 4.29.1%, n ) 3) for river water at both fortification levels. The values were much lower in the case of AMOX, especially in the river water samples, as shown in Table 2, whereas in the case of AMPI recoveries were always higher than 71% (RSD 7.2-9.7%, n ) 3) at both spiking levels. Recoveries on the order of 53.3% have been described in the literature for the analysis of PENV in water samples using reversed-phase silica and polymer-based sorbents29 and on the order of 36% for AMOX, 33% for PENG, and 58% for PENV in tap water using a mixed-phase sorbent30 and HPLCMS-MS detection. Other cartridges, such as Carbograph 4, showed recoveries around 76% for preconcentration of BLAs in water.31 For being a UV-based method, the results compare well with methods based on commercial SPE sorbents14 and demonstrate the broad applicability of this polymer for the determination of BLAs in aqueous samples. Finally, it is important to stress that the same cartridges were reused more than 100 times without losing their preconcentration ability. CONCLUSIONS This work demonstrates the applicability of a urea-based MIP for the preconcentration of eight β-lactam antibiotics in water samples. The optimized method is based on an SPE traceenrichment step using MIP cartridges followed by HPLC-UV (DAD) analysis without derivatization. The method has been successfully applied to the analysis of the antibiotics in tap and river waters and provides good recoveries and reproducibility. The cartridges can be reused for more than 100 assays without losing their preconcentration efficiency, which is promising for on-line preconcentration formats. Although the detection limits using the UV-based method are not very low, the optimized methodology can be easily applied with more sensitive instrumentation. This should be useful for a first screening of the selected β-lactam antibiotics in environmental water samples. ACKNOWLEDGMENT This work has been funded by the Madrid Community (Grant S-0505/AMB/0374), the European Regional Development Fund and the European Social Fund. J.L.U. thanks the Spanish Ministry of Science and Technology for a predoctoral grant. SUPPORTING INFORMATION AVAILABLE Graph of the eluting ability of different solvent mixtures. This material is available free of charge via the Internet at http:// pubs.acs.org. Received for review August 30, 2006. Accepted October 25, 2006. AC061622R

Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

701