Aminoglycosides (AGs) are classes of antibiotics most widely used in veterinary practice to treat bacterial infections of animals in livestock farming and bovine milk production, that causing residues in foodstuffs which produce adverse reaction in humans, as allergic reactions and antibiotic resistance[ ]
They are commonly used to treat foulbrood infection (caused by bacteria), and Nosema disease (caused by protozoa). Veterinary use of antibiotics is regulated by European Union (EU) and maximum residue limits (MRLs) have been established in different edible parts or products derived from animals, such as muscle, kidney, fat, liver, milk, and eggs [ ].The AGs have been widely used in animal husbandry for the treatment of infections caused by bacteria or growth promotion[ ]. Thus, the use of these antibiotics in dairy cattle may result in drug residues in their milk, especially if they are not used according to label directions [ , 10]. Aminoglycosides (AGs) are extremely hydrophilic compounds because having a lot of amino and hydroxyl groups in their chemical structures [ ].These drugs tend to accumulated in kidney as they are generally excreted via the urinary tract [10]. The toxicity of AGs includes (usually reversible) nephrotoxicity, (irreversible) ototoxicity in the cochlea and vestibular organs and, rarely, neuromuscular blockade and hypersensitivity reactions [ ]. Streptomycin and gentamicin are primarily vestibulotoxic, causing dizziness, ataxia, and/or nystagmus, whereas amikacin, neomycin, dihydrosterptomycin, and kanamycin are primarily cochleotoxic, causing permanent hearing loss [ ]. In this sense, European Governments are dedicating a lot of money in advertising campaigns and scientific projects to avoid the misapply of antibiotics in both animals and humans. Therefore, to ensure food safety, European Union (EU) has set maximum residue limits (MRLs) of antibiotics in foodstuffs of animal origin by means of the Commission Regulation 37/2010[2]. Consequently, there is a growing required to develop sensitive and reliable analytical methods, which monitor residues of trace-level AGs in complex matrices [ ]. AGs are difficult to analyze due to a lack of chromophore and poor retention on reversed-phase LC columns [ ] and due to their extreme polarity, AGs require various sample preparation and chromatographic conditions compared with other antibiotics [ ]. The critical challenge for a valid determination of trace-level AGs in high complicated matrices is related with the extraction and cleanup procedure. Various purification methods including solid-phase extraction (SPE) [ ] on-line SPE [ ] and matrix solid-phase dispersion extraction [ ].Other materials have higher selectivity, like molecularly imprinted polymers (MIPs) can promote cleaner extracts. MIPs are synthetic materials with artificially produced recognition sites capable of specifically catch target molecules [ ]. Therefore, the strong interaction between MIPs and target molecules makes them ideal for the selective extraction of compounds at trace levels, being of special interest for complex matrices [ ]. Because of AGs have a high polarity; they are not retained on the reversed-phase columns [8]. This fact presents a big analytical challenge. So, there were various methods to obtain some retention of the chromatographically unretained AGs using either ion pairing agents or hydrophilic interaction chromatography (HILIC) which is recently quoted by researchers but no explanation was found for their choice [ ].Sensitivity and selectivity in relation to problems have induced numerous authors to use liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) to analyze AGs [8, , ]
Because of these challenges, it is useful to consider alternative approaches to the chromatographic separation and subsequent MS analysis of these compounds. One chromatographic technique that has been investigated for the analysis of AG residues is the use of HILIC [ ]. HILIC–MS/MS to determine some aminoglycosides in serum has been proposed by Oertel et al [5]. This method has great advantages as compared with others, such a more volatile mobile phase and less ion suppression [ ]. Therefore the aim of proposed study was to develop a multi-residue method that would be simple and fast for routine regulatory analysis of the 11 aminoglycosides residues in different types of milk (whole cow’s milk, skimmed cow’s milk, goat milk, infant milk, milk containing omega3 and milk with isoflavon) using a recently commercially available MIPs. To the best of our knowledge, that is the first report about the use of MIPs combined with HILIC–MS/MS, demonstrating that both methodologies can be convenient and its potency for the determination of these antibiotics in milk samples.
2. Materials and methods
2.1. Reagents and materials
Due to the high absorption affinity of the AGs to polar surfaces and also their high photosensitivity, only laboratory polypropylene equipment was used during sample preparation, storage, and injection. Ultrapure water (Milli-Q Plus system, Millipore Bedford, MA, USA) was used throughout the work. Methanol (MeOH) and acetonitrile (MeCN) (LC-MS HiPerSolv grade) were supplied from VWR (Radnor, PA, USA). Formic acid, acetic acid and heptafluorobutyric acid (HFBA) were obtained from Sigma Aldrich (St. Louis, MO, USA). ammonium hydroxide (30%), potassium dihydrogen phosphate and dichloromethane were obtained from Panreac-Química (Barcelona, Spain). Ammonium acetate was obtained from Merck (Darmstadt, Germany). Vetranal grade analytical standards of Gentamicin (GENT), that was a mixture of GENT C1, GENT C1a and GENT C2, Apramycin (APM), Paromomycin (PRM), Dihydrostreptomycin (DHS), Spectinomycin (SPC), Kanamycin (KAM),Lincomycin (LIM), Amikacin (AM), Tobramycin (TOM) and Streptomycin (STP) were supplied by Fluka Analytical (Steinheim, Germany).
MIP extraction cartridges (SupelMIP AGs SPE Column, 50 mg, 3 mL) supplied by Supelco (Bellefonte, PA, USA) were used for extraction and clean-up process. Nylon syringe filters, 0.22 mm x 13 mm (Agela Technologies, New York, USA) were used for filtration of the sample extracts before injection into the LC-MS/MS system.
2.2 Preparing of standard solution
Individual stock standard solutions have been prepared by dissolving accurately weighed amounts in water to make a stock standard solution with 100 mg/L of concentration of each analyte and stored in a plastic tube in the dark at -20 °C. They were stable for at least 2 months. Tuning solution of each analyte (10 mg/L) was prepared by diluting the individual stock solution with the same solution. A working solution of a standard mixture has freshly prepared by proper dilution of the stock standard solutions with MeCN /H2O 25/75 in two groups (AM, APM, KAM, PRM, STP, TOM, GENT C1a, GENT C1 and GENT C2C2A) and (DHS and SPC) with 10 mg/L of concentration. In case of lower fortification mixture was required, an extra dilution of the AGs was prepared. These solutions were stored in plastic tubes at 2–4 °C and remained stable for up to 1 week.
2.3 Instrumentation
Separation was performed on an Agilent 1290 Infinity LC using a kinetex HILIC column (50 × 2.1 mm, 2.7 m) supplied by Phenomenex (Torrance, CA, USA). The mass-spectrometer measurements were performed on a triple quadrupole (QqQ) mass spectrometer API 3200 (AB Sciex, Darmstadt, Germany) with electrospray ionization (ESI). The instrumental data were collected using the Analysts Software version 1.5 with Schedule MRMTM Algorithm (AB Sciex). A centrifuge (Universal 320 model from Hettich, Leipzig, Germany), Avortex (Genie 2 model from Scientific Industries, Bohemia, NY, USA) and a pH meter has a resolution of ±0.01 pH unit (Crison model pH 2000, Barcelona, Spain) were used also during the sample preparation procedure.
2.4 UHPLC–MS/MS analysis
Separation was performed in a kinetex HILIC column (50 × 2.1 mm, 2.7µm) from Phenomenex (Torrance, CA, USA), using a mobile phase consisting of 150mM ammonium acetate containing 1% formic acid (solvent A), and MeCN (solvent B) at a flow rate of 0.5 mL min-1. The eluent gradient profile was as follow 80 % B at the beginning; change to 30% B in 2 min, change to 5% B in 4 min, change to 5 % B in 5 min and continue to 7 min and finally go back to the initial condition in 9 min and continues for 4 min. Under optimum conditions, all the analytes were eluted in 2.8 min, while the run time for each injection was 13 min. The temperature of the column was 35° C and the injection volume was 20 µL. The UHPLC system was coupled to a mass-spectrometer with ESI operating in positive ion mode, under the multiple-reaction monitoring (MRM) conditions. The ionization source parameters were: dry gas temperature, 700°C; curtain gas (nitrogen), 30 psi; ion spray voltage, 4000 V; collision gas, 5 and dry gas pressure (GS 1 and GS 2, both of them N2) 50 psi.
2.5 Sample treatment procedure
Samples of 2 g of whole cow’s milk sample (obtained from a local store) were spiked at different concentration levels using the working standard solutions of AGs. After spiking and homogenizing in vortex, 250 µL of trichloroacetic acid TCA 15 % were added for protein precipitation and homogenized and centrifuged at 9000rpm for 5 min. Then, for defatting sample, 1mL of n-hexan was added to the sample extract and mixing well, then, centrifuged at 9000rpm for 5 min. Then the upper layer which containing fat has been removed and the rest were diluted with approximately 2.5 mL of 50 mM potassium phosphate buffer pH 7.0, and shaken manually for 10s. The pH of the final solution was checked to be in rang 6.5 – 8.5, and if necessary it was adjusted with 30% ammonium hydroxide. Then, the final volume was adjusted to 5 mL with 50 mM potassium phosphate buffer pH 7.0. A 3 mL aliquot of this solution was loaded onto a SupelMIP AG SPE column (previously conditioned with 1mL of MeOH and 1mL of phosphate buffer pH 7.0) at a flow rate of approximately 0.2 mL min-1. After that, the cartridge washed with 3 mL of water at a flow rate lower than 0.5 mL min-1. Subsequently, strong vacuum (-20 in Hg) was applied for 5 min. Then, the MISPE cartridge was washed again with 1 mL of a mixture of dichloromethane: MeOH (50:50, v/v). After this washing step, a slight vacuum was applied for 10 s. Finally, the elution of the analytes was achieved using 1 mL of 1% formic acid in MeCN: H2O (20:80, v/v) with 5mM heptaflourobuteric acid (HFBA) and then eluted again with 2 ml of MeCN: H2O (20:80, v/v). This extract was filtered and injected directly in the UHPLC /MSMS.
3. Results and discussion
3.1 Optimization of chromatographic separation and MS/MS detection
Concerning the chromatographic conditions, aqueous standard solutions of AGs were used during the optimization of chromatographic separation. The mobile phase in HILIC separation plays important roles and its pH conditions are generally controlled by buffer solutions concentration. With respect to the MS sensitivity, the ammonium acetate was superior to the ammonium format [ ]. Therefore, the ammonium acetate has been used in different concentrations to obtain sharp peaks and in the same time to achieve good responses for all AGs. Different concentration of ammonium acetate (75-200mM) were tested, the concentration of 150 mM of ammonium acetate in the mobile phase was required to obtain sharp peaks and high intensity for all components. The use of acid in the mobile phase is required to improve the ionization step in ESI+. Therefore, different percentages of formic acid (0-2%) in solvent A were tested. Finally, with using a 1.5% formic acid obtaining high signal intensity, for this reason this was finally chosen for the rest of the experimental. Subsequently, the gradient was studied to get the best separation, peak shape and sensitivity in the shortest time. The flow rate was studied from 300 to 600 μl min-1 and finally 500 μl min-1 was selected as a compromise between signal, peak shape and run time. The column temperature of was studied between 25 °C and 55 °C and 35 °C was selected, as it provided the highest peak height with the best resolution and good analysis time. The injection volume was evaluated from 5 to 20 μl, and 20 μl was selected as optimum because with a lower injection volume the efficiency of the peaks was decreased.
After the chromatographic optimization, a study of the ionization source was carried out following the recommendations of the manufacturer: source temperature was tested between 300 and 750 °C and The study showed improvement on the signal in the studied AGs when the TEM was increased up 500 °C For this reason a TEM of 700 °C was selected as optimum.; curtain gas (nitrogen) was studied between 20and 35 psi and the best results were obtained with 30 psi as higher values produced a slightly decrease in the signals; ion spray voltage was evaluated from 3000 to 6000 V and finally 4000 V was selected as optimum and considered enough to obtain satisfactory signals.; and GAS 1 and GAS 2 (both of them nitrogen) were set to 50 psi.
3.2 Optimization of MISPE
Aminoglycosides antibiotics are extremely hydrophilic compounds that do not bind strongly to protein in the matrix [ ] .The use of MISPE has the possibility to simplify the extraction of AGs from a complicated matrix such as milk, as long as a higher selectivity and a lower sample manipulation. Initially, the protocol proposed by Supelco for the SupelMIP AGs SPE Columns for the determination of Neomycin, GENT C1, DHS, STP, Geneticin (G418-2), AM, TOM, KAM, APM, Hygromycin, Puromycin and SPC in honey samples was followed with some modification.
Referring to the precipitation of milk proteins, different authors have been used TCA as a participating agent [ , , , ]. Kaufmann and Maden [11] discussed recovery efficiency be using different concentration of TCA in extraction treatment. To achieve a plateau for 11 AGs, 1mL of 15%, 25% and 50% TCA were tested to precipitate protein of 2 ml milk. And finally 15 % of TCA were enough to precipitate milk protein. Then different volumes of TCA 15% (250 – 1000µl) were evaluated to reach the highest recovery, whereas the recovery % for all analytes were increased by decreasing the volume of TCA, in the same time, the efficiency of the precipitation of milk proteins was the same with all of them. Thus, 250 µL TCA 15 % was applied to extract multianalytes and eliminate the turbidity with prominent effectiveness of protein precipitation.
Referring to the washing steps, the step of washing with 1mL of 0.1 % ammonium hydroxide tested to use it or removing from the schedule, whereas it was removing the recovery % has been increased for most analytes. That might be because the AGs were lost in this washing step.
On the other hand, the washing with 1 mL of 40:60 MeCN: H2O (v/v) was evaluated if some AGs were lost in this step or not by collecting and analyzing this wash step. However the results obtained showed that in this step about 27.4, 30.6, and 39.3 and 100 % of AM, APM, KAM and LIM, respectively, were lost. In the same time different percentage of MeCN: H2O (40:60, 50:50 and 60:40) were tested to obtain the highest recovery, but without using this washing step were obtained the highest recovery.
Referring to elution solvent, because the polarity dependence of eluent, ion pair reagent HFBA played a key role in the elution procedure [8]. Different concentration of HFBA (0 – 10mM) were tested in the elution solvent (1% formic acid in MeCN: H2O (20:80, v/v)). By using 5mM of HFBA were obtained the highest percentage of recovery for all analytes. On the other hand, for obtaining good peak shape, 1 mL of 1% formic acid in MeCN: H2O (20:80, v/v) with 5mM of HFBA were used at first to elute the AGs from the column and then diluted with 2 mL of 1% formic acid in MeCN: H2O (20:80, v/v).
3.3 Method Validation
The proposed method for the determination of AGs in milk was characterized in terms of linear dynamic ranges, limits of detection (LODs) and limits of quantification (LOQs), precision and trueness using whole cow’s milk as representative matrix.
3.4 Calibration curves and analytical performance characteristics
Calibration curves were assessed by spiking blank samples of whole milk (50, 100, 150, 200 and 250 µg kg−1) for AM , APM, KAM, PRM, STP, TOM, GENT C1a, GENT C1, GENT C2C2A and (10, 25, 50, 100 and 150 µg kg−1) for DHS and SPC before the extraction process. Each sample was prepared following the proposed method and injected in triplicate.
LODs and LOQs were calculated as 3×S/N and 10×S/N, respectively. As can be seen in Table 1, LOQs were always lower than the reference MRLs. Therefore, the proposed method could be used for the determination of these compounds in the selected matrices at very low concentration levels.
3.5 Precision study
Both repeatability and intermediate precision were tested by application of the proposed MISPE-UHPLC-MS/MS method in whole milk samples spiked at two different concentration levels of AGs (50 and 100 µg kg−1 for AM, APM, KAM, PRM, STP, TOM, GENT C1a, GENT C1, GENT C2C2A and for DHS and SPC were 10 and 25 µg kg−1). To check the repeatability (intraday precision), three samples were prepared and injected in triplicate on the same day, under the same conditions. Similar procedure was carried out in the evaluation of intermediate precision (interday precision). Thus, during three consecutive days, one sample per day was prepared and injected in triplicate. The results, expressed as %RSD of peak areas, are shown in Table 2. Good precision (RSD lower than 13%) was obtained in all cases. So, it could be concluded that the obtained results are in agreement with the current demand [ ].
3.6 Trueness assessment
The trueness of the proposed method was assessed by recovery studies in different types of milk samples (whole cow’s milk, skimmed cow’s milk, goat milk, infant milk, milk containing omega3 and milk with isoflavon) spiked at two different concentration levels of each AGs (50 and 100 µg kg−1 for AM, APM, KAM, PRM, STP, TOM, GENT C1a, GENT C1, GENT C2C2A and for DHS and SPC were 10 and 25 µg kg−1).
The absolute recoveries have been calculated by comparing concentration of AGs in milk samples spiked before the MISPE procedure with concentration in extracts of milk samples spiked after the MISPE procedure using calibration curve. Each sample was analysed in triplicate and injected three times. Blank samples were previously analysed to check the presence of AGs; none of them gave a result above the LOQs of the method. The recoveries of (whole cow’s milk, skimmed cow’s milk, goat milk, infant milk and milk with isoflavon) were between 70% and 106% for all analytes except TOM it was between 45 to 65, with satisfactory precisions for all analytes (see Table 3), fulfilling current legislation [21]. Whereas, with sample of milk containing Omega 3, the recoveries of (APM, GENT C1a, KAM, PRM, TOM) were lower than 70 %, that may be because the chemical composition of this type of product is completely different than other type of tested milk.
A typical extracted ion chromatogram corresponding to a whole cow’s milk sample spiked with 25µgkg-1 for DHS and SPC and 100 µg kg-1 for AM , APM, KAM, PRM, STP, TOM, GENT C1a, GENT C1, GENT C2C2A and analysed by the proposed MISPE–UHPLC–MS/MS method is shown in Fig.1.
4. Conclusions
In the present study, UHPLC-MS/MS combined with MISPE has indicated to be a reliable and powerful method for the simultaneous quantification and confirmation of 11 aminoglycosides in different types of milk (whole cow’s milk, skimmed cow’s milk, goat milk, infant milk, milk containing omega3 and milk with isoflavon). A simple, selective and high sensitive designing for the determination of these AGs has been developed and validated. The results showed that MISPE as a robust tool for extraction and sample clean-up, eliminating complicated steps. Calibration curves were established in the presence of matrix and the low LOQs obtained allowed determining the 11 AGs at concentrations lower than the limits established by current legislation for AGs in milk, with satisfactory precisions. In addition, trueness has been successfully evaluated for whole cow’s milk, skimmed cow’s milk, goat milk, infant milk, milk containing omega3 and milk with isoflavon, obtaining good recoveries for all AGs, except TOM in tested samples and (APM, GENT C1a, KAM, PRM, TOM) in milk containing omega 3 due to dissimilar chemical composition. The developed method is simple, rapid, low solvent consumption and inexpensive providing good sensitivity and selectivity. Thus, these results showed the suitability of this MISPE-UHPLC-MS/MS procedure for the monitoring of AGs residues in less explored samples, as milk.
Acknowledgment
AMH thanks the “Erasmus Mundus – Al Idrisi II program” for a doctoral grant. The authors gratefully acknowledge the financial support of the Junta de Andalucía (Excellence Project Ref: P12-AGR-1647).