Anal. Chem. 2001, 73, 1614-1621
Simultaneous Determination of Five β-Lactam Antibiotics in Bovine Milk Using Liquid Chromatography Coupled with Electrospray Ionization Tandem Mass Spectrometry Sonja Riediker and Richard H. Stadler*
Nestle´ Research Center, Nestec Ltd., Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland
A multiresidue method for the detection of five important β-lactam antibiotics (amoxicillin, ampicillin, cloxacillin, oxacillin, penicillin G) in fresh milk is presented that allows quantitation of the analytes well below established legislative limits. The method avoids the use of acid during the extraction procedure and entails a cleanup step over a C18 cartridge. The analytes are separated and detected by liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) using a stable isotope-labeled internal standard. Mass spectral acquisition is done in the positive ion mode by applying selected reaction monitoring of two or more fragmentation transitions per analyte to provide a high degree of sensitivity and specificity. The typical recoveries for all five β-lactams in fresh milk ranged from 76 to 94% at a fortification level of 4 µg/kg. This study also addresses common problems encountered in the stability of penicillins during sample preparation as well as the employment of postcolumn infusion of a standard compound to verify potential matrix-induced signal suppression in ESI-MS. Antibiotics of the β-lactam group are part of a wide variety of antimicrobial agents that are used to treat infectious human and animal diseases and to enhance growth and yield in agriculture (livestock and crop production, fish farming). The widespread and intense employment of antibiotics has led to an increasingly potential risk for human and animal health mainly because of hypersensitivity of some individuals to the parent compounds and/ or their metabolites and increasing development of resistance against antibiotics in certain pathogenic bacteria.1,2 In fact, stringent quality control procedures are required to protect consumers and the European Union (EU) has established maximum residue limits (MRLs) for antibiotics in food. For example, the MRLs for the major β-lactams amoxicillin (AMOX), ampicillin (AMPI), and penicillin G (PEN G) in bovine milk are set at 4 µg/kg, and at 30 µg/kg for oxacillin (OXA) and cloxacillin (CLOX).3 * Corresponding author: (e-mail)
[email protected]; (fax) +41 21 785 85 53; (phone) +41 21 785 83 60. (1) Aiello, S. E. The Merck Veterinary Manual, 8th ed.; Merck & Co., Inc.: Whitehouse Station, NJ, 1998; p 1745. (2) Schwarz, S.; Noble, W. C. Vet. Dermatol. 1999, 10, 163-176.
1614 Analytical Chemistry, Vol. 73, No. 7, April 1, 2001
To ensure that legislation is respected at the farming level, commercial screening tests such as microbial inhibition tests (fourplate test, Delvotest P), receptor assays (Charm test), and immunoassays (LacTek-Bl) are commonly employed. However, these tests do not permit compound-specific identification and quantitation, and false-positive results may be generated by, for example, the presence of natural microbial inhibitors or the greater stability of some β-lactam antibiotics (cloxacillin or oxacillin) toward penicillinase.4 Consequently, sensitive analytical methods are required for confirmation and quantitation of antibiotic residues. The determination of β-lactam antibiotics at trace levels in fresh milk requires a sensitive analytical method that in particular addresses the impact of sample preparation and the analytical step on the stability of the analytes (e.g., pH, organic solvents).1,5,6 This may be the reason some multiresidue methods described in the literature differ significantly in sample preparation, while the determinative step is usually based on reversed-phase liquid chromatography (LC) coupled with UV detection.4,7-12 Furthermore, many of the methods described are tedious, requiring a derivatization step and/or use an acidification step during sample preparation which is critical with regard to analyte stability. A number of authors have reported multiresidue methods13-19 for the determination of β-lactams in milk by LC-MS techniques, but only a few of these methods have been developed that can detect the penicillins at required regulatory limits.13,16 Straub et al.14 described a short sample preparation based on compound extraction using acetonitrile/water (1:1) followed by ultrafiltration using a cutoff filter of 10 000 Da and subsequent LC-ESI-MS analysis. This method was considered as fast but problematic in truly routine operations, especially with regard to the high injection volumes required to achieve low detection limits.15 (3) Council Regulation (EEC) No. 2377/90 of 26 June 1990, OJ No L 224, 18.08.1990. (4) Fletouris, D. J.; Psomas, J. E.; Mantis, A. J. J. Agric. Food Chem. 1992, 40, 617-621. (5) Schwartz, M. J. Pharm. Sci. 1965, 54 (3), 472-473. (6) Tyczkowska, K. L.; Voyksner, R. D.; Aronson, A. L. J. Chromatogr. 1992, 594, 195-201. (7) Moats, A. W.; Harik-Khan, R. J. AOAC Int. 1995, 78, 49-54. (8) Moats, A. W.; Romanowski, R. D. J. Chromatogr., A 1998, 812, 237-247. (9) Edder, P.; Cominoli, A.; Corvi, C. Mitt. Lebensm. Hyg. 1999, 90, 291-304. (10) Sørensen, L. K.; Rasmussen, B. M.; Boison, J. O.; Keng, L. J. Chromatogr., B 1997, 694, 383-391. (11) Ibach, A.; Petz, M. Z Lebensm. Unters. Forsch. A 1998, 207, 170-173. (12) Verdon, E.; Couedor, P. J. Chromatogr., B 1998, 705, 71-78. 10.1021/ac0011383 CCC: $20.00
© 2001 American Chemical Society Published on Web 03/06/2001
Furthermore, many methods do not report the use of appropriate internal standards, which are important for reliable quantitation and method validation. The use of nafcillin and penicillin V as internal standards in milk samples has been reported and the methods have been validated.13,16,19 However, nafcillin and penicillin V are representatives of the β-lactam family and may also be applied in veterinary medicine, which questions the suitability of these compounds as valid internal standards. A further important point not adequately addressed in some methods for determining β-lactam antibiotics in milk is the issue of analyte stability, both in the sample extracts and during the analytical step (e.g., overnight runs). In this study, we elaborate on the common problems encountered in analyte stability during the extraction, cleanup, and determinative steps. Moreover, the method presented is quantitative for the simultaneous analysis of both monobasic and amphoteric β-lactams at trace levels in fresh milk and can be employed as a confirmatory quality control tool for milk samples that provide ambiguous results in rapid screen tests. EXPERIMENTAL SECTION Materials and Reagents. The β-lactam antibiotics amoxicillin (13.7% H2O), ampicillin (sodium salt), penicillin G (sodium salt), azlocillin (sodium salt), and cloxacillin (sodium monohydrate salt) were purchased from Sigma (Buchs, Switzerland). Oxacillin (sodium monohydrate salt) was obtained from Riedl-de Hae¨n GmbH & Co (Seelze, Germany). Potassium benzyl(phenyl-d7) penicillate (chemical purity >95%, d7-PEN) that was used as internal standard (IS) was custom synthesized by Toronto Research Chemicals Inc. (North York, ON, Canada). The antibiotics were stored in a dried atmosphere at 4 °C (AMOX, AMPI, CLOX, OXA), at ambient temperature (PEN G, azlocillin), or at -20 °C (IS). As recommended by Tyczkowska et al.,6 stock solutions (1 mg/mL) were prepared with a mixture of ethanol, acetonitrile, and water (1:1:2, v/v) and stored -20 °C for one month. Immediately prior to fortifying the samples, working standards were prepared by diluting the stock solutions with distilled water. Formic acid, n-hexane, ethanol, and sodium chloride were of p.a. grade (Merck, Darmstadt, Germany). Water was either purified in-house using a Bu¨chi Fontavapor 260 (Flawil, Switzerland) or purchased from Merck (HPLC grade, LiChrosolv). Methanol (LiChrosolv, Merck) and acetonitrile (J. T. Baker, Phillipsburg NJ) were of HPLC grade. Phosphate buffer (100 mM) was prepared by dissolving 4.45 g of disodium hydrogen phosphate dihydrate (p.a. grade, Merck) in 250 mL of distilled water and sonifying to dissolve. Bakerbond C18 solid-phase extraction cartridges (3 mL, 500 mg, SPE) were obtained from J. T. Baker. (13) Blanchflower, W. J.; Hewitt, S. A.; Kennedy, D. G. Analyst 1994, 119, 25952601. (14) Straub, R.; Linder, M.; Voyksner, R. D. Anal. Chem. 1994, 66, 3651-3658. (15) Heller, D. N.; Ngoh, M. A. Rapid Commun. Mass Spectrom. 1998, 12, 2031-2040. (16) Daeseleire, E.; De Ruyck, H.; Van Renterghem Conference Proceeding, EuroResidue IV, Conference on Residues of Veterinary Drugs in Food; Veldhoven, NL, 2000; pp 333-338. (17) Tyczkowska, K. L.; Voyksner, R. D.; Straub, R. F.; Aronson, A. L. J. AOAC Int. 1994, 77, 1122-1131. (18) Parker, C. E.; Perkins, J. R.; Tomer, K. B. J. Chromatogr. 1993, 616, 4557. (19) Boison, O. K.; Keng, L. J.-Y.; MacNeil, J. D. J. AOAC Int. 1994, 77 (3), 565-570.
Instrumentation. Measurements were conducted on a Quattro LC triple quadrupole mass spectrometer (Micromass, Manchester, U.K.) that was equipped with a Z-spray electrospray ion source and coupled with an Alliance liquid chromatograph 2690 (Waters, Rupperswil, Switzerland). A Labpro six-port two-position flow processor (PR700-100, Rheodyne, Cotati, CA) directed the LC effluent to the MS only during analyte separation. Instrument control, data acquisition, and evaluation were done by running the software MassLynx version 3.4. A PEEK precolumn filter (2µm PEEK polymer frit) from Upchurch Scientific (Oak Harbor, WA) was installed to protect the separation column. Milk Samples. Pasteurized full-cream milk samples of different brands were purchased off the shelf and stored at 4 °C. A raw milk sample with naturally incurred residues (positive response with the rapid immunoassay-based screen tests) was obtained from a milk collection center in the United Kingdom and shipped in a frozen state to the Nestle´ Research Center in Switzerland, where the sample was stored at -30 °C until analysis. Milk Extraction. A portion of the milk sample (20 mL) was weighed in a 40-mL centrifugation tube (Teflon). The internal standard was added to achieve a final concentration of 10 µg/kg d7-PEN. The sample was vortex-mixed for 1 min and then centrifuged at 13800g in an Avanti J-25I centrifuge (Beckman Instruments Inc., Fullerton, CA) for 30 min (3-4 °C). After centrifugation, an aliquot (12 mL) of the defatted milk was transferred to centrifuge tubes containing the same volume of 0.1 M phosphate buffer (pH 9.2). This mixture was vortex-mixed for 1 min. Subsequently, 12 mL of n-hexane was added and the sample was shaken mechanically (KS 501 D shaker, IKA-Labortechnik, Staufen i. Br., Germany) for 10 min. The mixture was then centrifuged at 2790g for 10 min at 9-10 °C. The organic and emulsion layers were carefully removed by vacuum suction and an aliquot (15 mL) of the aqueous extraction phase was loaded onto a SPE cartridge (Bakerbond C18, 500 mg) preconditioned consecutively with each 10 mL of methanol, distilled water, and finally 5 mL of a 2% NaCl solution. After sample penetration (flow rate ∼0.25 mL/min), the sorbent was rinsed rapidly with 2 mL each of 2% NaCl and distilled water and suctioned to “dryness” for 5 min. The analytes were eluted (1.5 mL) with a solvent mix made up of 50 mM phosphate buffer (pH 8.0, adjusted with 1% formic acid) and acetonitrile (1:1, v/v), prepared directly before use. The effluent was evaporated to a volume of
