VEACH
ET AL.: JOURNAL OF
AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017
1
VETERINARY DRUG RESIDUES
Quantitation and Confirmation of Chloramphenicol, Florfenicol, and Nitrofuran Metabolites in Honey Using LC-MS/MS BRIAN T. VEACH, RENEA ANGLIN, THILAK K. MUDALIGE, and PAULA J. BARNES U.S. Food and Drug Administration, Arkansas Laboratory, Office of Regulatory Affairs, 3900 NCTR Rd, Jefferson, AR 72079
This paper describes a rapid and robust method utilizing a single liquid–liquid extraction for the quantitation and confirmation of chloramphenicol, florfenicol, and nitrofuran metabolites in honey. This methodology combines two previous extraction methods into a single extraction procedure and utilizes matrix-matched calibration standards and stable isotopically labeled standards to improve quantitation. The combined extraction procedure reduces the average extraction time by >50% when compared with previously used procedures. The drug residues were determined using two separate LC-tandem MS conditions. Validation of all the analytes was performed, with average quantitation ranging from 92 to 105% for all analytes and the RSDs for all analytes being ≤12%.
H
oney bees have a significant economic impact across the world. In the United States alone, it is estimated that more than $15 billion in increased crop value year-peryear is directly associated with honey bees. It has been shown that, on average, one in every three mouthfuls of our diet is associated directly or indirectly with honey bee pollination; furthermore, honey bees produce more than 140 million pounds of honey in the United States annually (1–3). However, in recent years, there have been reports of declining numbers of honey bees (2–6). The U.S. Department of Agriculture published a document in 2013 that the beekeeping industry is reporting a high percentage of losses each year to parasites and pests, poor nutrition, sublethal exposure to pesticides, and pathogens (7). There are several bee diseases that have contributed to the declining bee colony populations. Three of the most commonly referenced diseases are American foulbrood (AFB), European foulbrood, and Nosema disease. AFB disease is caused by a spore-forming bacterium. It is one of the most destructive and Received June 26, 2017. Accepted by JB October 24, 2017. The views expressed in this paper are those of the researchers and should not be interpreted as the official opinion or policy of the U.S. Food and Drug Administration, Department of Health and Human Services, or any other agency or component of the U.S. government. The mention of trades names, commercial products, or organizations is for clarification of the methods used and should not be interpreted as an endorsement of a product or manufacturer. Corresponding author’s e-mail:
[email protected] Supplemental Information is available on the J. AOAC Int. Web site, http://aoac.publisher.ingentaconnect.com/content/aoac/jaoac DOI: https://doi.org/10.5740/jaoacint.17-0262
widespread honey bee brood diseases. European foulbrood is another disease that is closely related to AFB, but the causative bacterium is different and does not form spores. It is common in the spring when brood rearing is at its height. Nosema disease, or nosemosis, is a parasitic disease of adult honey bees caused by two species of microsporidia, Nosema apis and N. ceranae, which form spores. Many believe that Nosema disease may be the most damaging adult bee disease. The disease directly infects the epithelial cells of the hind gut (ventridulus) of the digestive tract in the adult bee (8–11). Although these are three of the more commonly referenced diseases, there are a host of other diseases that infect bee colonies, and several of these diseases are treated through the use of antimicrobial drugs. Chloramphenicol (CAP), florfenicol (FF), and nitrofurans (NFs; nitrofurantoin, furazolidone, nitrofurazone, and furaltadone) are antimicrobial drugs that are sometimes used in the treatment of bacteria in food-producing animals. CAP and FF are used as an agent against both Gram-positive and Gram-negative bacteria (12–15). NFs are used for the treatment of bacterial and protozoan infections (15–21). Due to the apparent effectiveness and availability of these veterinary drugs, they are sometimes used to treat bee colonies. However, serious health risks are associated with these compounds and their metabolites. These risks include, but are not limited to, aplastic anemia, bone marrow suppression, increased risk of leukemia, and the potential for carcinogenic and mutagenic effects (12–22). NFs are administered in an inactive form; however, they are converted to active forms through normal metabolic processes. The NFs are rapidly metabolized to the following active forms: nitrofurantoin into 1-aminohydantoin (AHD), nitrofurazone into semicarbazide (SC), furaltadone into 3-amino-5-morpholinomethyl-2-oxazolidinone (AMOZ), and furazolidone into 3-amino-2-oxazolidinone (AOZ; 15–19). Because of the severe health risks associated with these drugs, their usage is not allowed in any food-producing animals in several countries, including the United States and the European Union (EU). However, the fact that they are effective and readily available in some countries makes the probability of their use a concern for many countries. Therefore, the EU and the United States have set minimum required performance limits (MRPLs) in honey for CAP at 0.3 µg/kg and NF metabolites at 1 µg/kg (23). The United States has additionally set an MRPL for FF in honey at 5 µg/kg. It should be noted that the MRPL does not indicate that concentrations below this level are safe or permitted, but is only guidance to the analyst that methods of use must be able to accurately quantitate and confirm residues at these levels. The vast majority of quantitative and confirmatory methods available for the detection of NFs are targeted toward seafood, and the few methods that are specific to honey are not inclusive
2 VEACH ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 of CAP and FF. Therefore, in order to encompass all of these residues, it is necessary to perform a minimum of two separate extractions. In order to increase throughput and efficiency, research efforts at the U.S. Food and Drug Administration (FDA), Arkansas Laboratory, have focused on the development of a method that can extract all of the aforementioned residues and analyze them via LC-tandem MS (MS/MS). Negative electrospray ionization (ESI) was used for CAP and FF analysis and positive atmospheric pressure chemical ionization (APCI) for the NF metabolites. Experimental Apparatus
(n) CAP reference standard.—Sigma-Aldrich (Part No. C0378). (o) CAP-d5 internal standard.—Cambridge Isotopes (Part No. DLM-1190-0; Andover, MA). (p) SCA-13C-15N2 internal standard.—Sigma-Aldrich (Part No. 338882). (q) AMOZ-d5 internal standard.—Sigma-Aldrich (Part No. 33881). (r) AHD-13C3 internal standard.—Santa Cruz Biotechnology (Part No. 208596; Santa Cruz, CA). (s) AOZ-d4 internal standard.—Santa Cruz Biotechnology (Part No. 216397).
Standards Preparation
(a) LC-MS/MS.—AB SCIEX Qtrap 5500, with Analyst software (Framingham, MA). (b) HPLC.—Agilent Technologies, 1260 series (Santa Clara, CA). (c) Chromatographic column.—Agilent Technologies Zorbax Eclipse XDB-C18 column (4.6 × 50 mm, 1.8 µm). (d) Centrifuge.—Thermo Fisher, capable of 3700 × g (Milford, MA) (e) Autosampler vials with inserts.—Agilent Technologies, 2 mL (Part No. 5188-8594). (f ) Shaking water bath.—Julabo SW22 (Allentown, PA). (g) PTFE syringe filter.—Fisher Scientific; 13 mm, 0.2 µm (Part No. 9720002; Houston, TX). Standards and Reagents (a) Acetonitrile.—LC-MS grade. (b) Methanol.—LC-MS grade. (c) Water.—LC-MS grade or equivalent. (d) Formic acid.—LC-MS grade. (e) Ammonium acetate.—LC-MS grade. (f ) Ethyl acetate.—HPLC grade. (g) Methanol.—LC-MS grade. (h) 0.125 M hydrochloric acid (HCl).—Reagent grade. (i) Dipotassium hydrogen phosphate (K2HPO4).—Reagent grade. ( j) Nitrofurazone reference standard.—Sigma-Aldrich (Part No. PHR1196; St. Louis, MO). (k) Nitrofurantoin reference standard.—Sigma-Aldrich (Part No. 1464001). (l) Furaltadone reference standard.—Sigma-Aldrich (Part No. F9130). (m) Furazolidone reference standard.—Sigma-Aldrich (Part No. F9505).
NF analytical and internal standard stock solutions (100 µg/mL) were prepared in a methanol–water solution (80 + 20, v/v). CAP stock and internal standard stock solutions (20.0 µg/mL) were prepared in methanol. FF stock and internal standard stock solutions (100 µg/mL) were prepared in methanol. A mixed intermediate analytical standard was then prepared at a concentration of 300 ng/mL for FF, 40.0 ng/g for each of the NFs, and 20.0 ng/g for CAP in a methanol–water solution (80 + 20, v/v). Additionally, a mixed intermediate internal standard solution was prepared at a concentration of 300 ng/mL for deuterated FF, 40.0 ng/g for each of the deuterated or isotopically labeled NFs, and 20.0 ng/g for the deuterated CAP standard in a methanol–water solution (80 + 20, v/v). Five extracted calibration standards were prepared by adding approximately 3.00 g ± 0.03 control matrix to a 50 mL polypropylene centrifuge tube. Each tube was fortified with the mixed intermediate analytical standard (see Table 1). The calibration standards were additionally fortified with 150 µL mixed intermediate internal standard solution (FF-d4, 15.0 ng/g; stable isotopically labeled NF metabolites, 2.00 ng/g; and CAP-d5, 1.00 ng/g).
Sample Preparation A volume of 150 µL mixed internal standard solution was added to a 3.00 g ± 0.03 negative control honey matrix in a 50 mL polypropylene centrifuge tube. Approximately 10 mL of 0.125 M HCl and 200 µL of 100 mM 2-nitrobenzaldehyde were added to each centrifuge tube. Each tube was mixed on a vortex mixer/shaken (approximately 10 min) and placed in the SW22 Julabo shaking water bath for the acid hydrolysis and derivatization process. The bath was set to approximately 37°C at approximately 80 rpm for 16 h.
Table 1. Calibration standards and concentrations levels Metabolite conc., ng/g Calibration standard
Intermediate standard volume, µL
Nitrofuran
Chloramphenicol
Florfenicol
1
45.0
0.600
0.300
4.50
2
75.0
1.00
0.500
7.50
3
150
2.00
1.00
15.0
4
225
3.00
1.50
22.5
5
450
6.00
3.00
45.0
VEACH
ET AL.: JOURNAL OF
AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017
3
Table 2. MS/MS acquisition parameters for the nitrofuran metabolite analysis on the AB SCIEX Qtrap 5500 using APCI Nitrofuran metabolite
Precursor ion, m/z
AHD
AOZ
AMOZ
SC
249.0
236.0
335.1
209.1
CE, Va
Product ions, m/z c
17
104.0
30
134.0
c
178.1
22
134.0c
18c
104.0
29
149.0
20
291.2c
16c
262.2
23
128.0
30
166.1c
14c
134.0
15
DP, Vb
Retention time, min
120
2.5
120
2.6
70
1.7
80
2.7
2.5
192.1
15
AHD-13C3
252.0
134.0c
16c
70
AMOZ-d5
340.1
296.2c
16c
70
1.6
AOZ-d4
240.0
179.1c
20c
120
2.5
SC-13C-15N2
212.1
168.1c
14c
80
2.7
a
CE = Collision energy.
b
DP = Declustering potential.
c
Product ions used for quantitation.
The tubes were removed upon completion of the acid hydrolysis and derivatization step, and 5 mL of 1 M K2HPO4 were added to each tube to adjust the pH to 7.3 ± 0.3. The pH can be further adjusted, if needed, with diluted NaOH or HCl. Approximately 2 g NaCl and 10 mL ethyl acetate were added to each tube, followed by mixing on the vortex mixer/shaking for approximately 10 min. The contents were centrifuged for 10 min at 3700 × g (Thermo Fisher), and the upper organic layer was transferred to a 15 mL polypropylene centrifuge tube and evaporated to dryness at 50°C. Samples were reconstituted with 250 µL reconstitution solution (40% methanol and 60% of 8.5 mM ammonium acetate in 0.1% formic acid). Reconstituted samples were mixed on the vortex mixer (approximately 10 s), sonicated (5 min), and filtered through a 0.2 µm PTFE syringe filter (Pall Life Science, Port Washington, NY) into a 2 mL autosampler vial with insert (Agilent Technologies). LC-MS/MS The prepared extracts were analyzed under two separate LCMS/MS conditions. NF metabolites were assayed using the Agilent 1260 HPLC system coupled to the AB SCIEX Qtrap 5500 mass spectrometer using APCI. Ten-microliter samples were injected into the XDB-C18 reversed-phase column (50 × 4.6 mm, 1.8 µm; Agilent Zorbax Eclipse) and eluted with 8.5 mM ammonium acetate in 0.1% formic acid (NF mobile phase A) and methanol (NF mobile phase B), followed by column cleaning and regeneration. NF mobile phase A was held at 60% (600 µL/min) for 2.3 min. At 2.4 min, NF mobile phase B (600 µL/min) was ramped up to 80% for 1 min. At 3.5 min, NF mobile phase B was ramped down to 40% (750 µL/min) and held for 1.5 min. Data from the mass spectrometer was acquired utilizing scheduled multiple-reaction monitoring (MRM) mode in positive APCI
(precursor ions and MS/MS fragment ions are listed in Table 2). The nebulizer current was optimized at 5 µA and the source heater at 300°C. CAP and FF were analyzed using the Agilent 1260 HPLC system coupled to the AB SCIEX Qtrap 5500 mass spectrometer using ESI. Two-microliter samples were injected into the XDBC18 reversed-phase column (50 × 4.6 mm, × 1.8 µm; Agilent Zorbax Eclipse). CAP and FF were eluted from the column with a gradient flow (750 µL/min) of water (mobile phase A) and acetonitrile (mobile phase B). Mobile phase A was held at an initial time of 0.5 min at 75%; subsequently, the proportion of mobile phase B was increased linearly to 50% in the following 3 min. At 3.6 min, the proportion of mobile phase B was increased to 90%, followed by a hold of 1 min at 90%. At 4.7 min, mobile phase B was returned to 25% and re-equilibrated for 1.3 min prior to the subsequent injection. Mass spectrometer data were
Table 3. MS/MS acquisition parameters for the chloramphenicol analysis on the AB SCIEX Qtrap 5500 using ESI
Analyte CAP
Precursor ion, m/z
Product ions, m/z
CE, Va
Retention time, min
321.1
151.8b
–23b
2.9
193.8
–16
257.2
–16
CAP-d5
326.1
262.2b
–16b
2.9
FF
356.0
185.0b
–38b
2.7
119.0
–35
335.8
–20
188.0b
–7b
FF-d3
359.0
a
CE = Collision energy.
b
Product ions used for quantitation.
2.7
4 VEACH ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017
Figure 1. LC-MS/MS nitrofuran metabolite quantitation ion chromatograms of a 0.600 ng/g extracted fortified control honey.
collected in negative ESI mode using scheduled MRM (the precursor and MS/MS fragment ions obtained are listed in Table 3). The spray voltage was optimized at –3000 V, with a sheath gas temperature at 400°C. The declustering potential was set to –70 V for all analytes. Data Analysis Quantitation was performed for each analyte of interest by calculating the ratio of the chromatographic area of the quantitation ion with respect to the chromatographic area
(quantitation ion) of the following stable isotopically labeled internal standards: AMOZ to AMOZ-d5, SC to SC-13C-15N2, AOZ to AOZ-d4, AHD to AHD-13C3, FF to FF-d3, and CAP to CAP-d5 (see Figures 1 and 2). Each representative ratio was plotted against the concentration of the corresponding matrixmatched calibration standard. The linear calibration curve fit obtained yielded a regression, R2, of ≥0.995. For positive confirmation, all product ions must be detected, the associated chromatographic peak must exhibit a retention time (RT) within 5% of the average RT of the calibration standards, and the product ion ratios must be within 20% of the product ion ratios obtained from the calibration standards (24, 25).
VEACH
ET AL.: JOURNAL OF
AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017
5
Figure 2. LC-MS/MS chloramphenicol quantitation ion chromatogram of a 0.300 ng/g extracted fortified control honey and the florfenicol quantitation ion chromatogram of a 4.50 ng/g extracted fortified control honey.
multiplying the SD by 10. The recovery data for the LOQs and MDLs are shown in Tables 4–6; the calculated MDLs and LOQs are presented in Table 7.
Analysis of Reference Materials and Commercial Products Reference materials were obtained from commercially available sources and were prepared as described in Sample Preparation. Samples were quantitated using matrix-matched standards that were previously screened and determined to be free of FF, CAP, and NF metabolites of interest using external methods (14–17). LOD and LOQ Studies The method detection limits (MDLs) and LOQs for each analyte were determined on the basis of replicate analysis (n = 7). The MDL of each analyte was calculated by multiplying the SD by Student’s t-value at the 99% confidence level (3.143) and the LOQ was calculated by
Results and Discussion Method Optimization Method optimization consisted of a 2-fold process. The first step was to develop instrument methods that would provide optimal sensitivity and chromatography. Subsequently, upon completion of instrument optimization, a new extraction method was needed to encompass all the drug residues of interest. Although previous research efforts have demonstrated CAP and its internal standard can be detected using APCI and NF metabolites can be assayed using ESI, optimal detection limits could not be achieved without utilizing both ESI and APCI (11).
Table 4. Average recoveries and RSDs of the nitrofuran metabolites in honey (n = number of replicates)a Fortification level n=7 Nitrofuran metabolite
0.600 ng/g
n=9
n=9
n = 10
n=1 12.0 ng/g
1.00 ng/g
2.00 ng/g
3.00 ng/g
AHD
98.6% (3.2%)
b
92.3% (7.9%)
94.9% (11%)
92.8% (12%)
105%
AMOZ
96.3% (3.4%)b
98.2% (3.8%)
103% (2.8%)
101% (3.2%)
94.2%
AOZ
96.0% (5.4%)b
98.9% (6.6%)
99.7% (2.9%)
97.3% (3.7%)
97.5%
SC
98.9% (3.2%)b
99.4% (4.4%)
96.9% (3.9%)
97.0% (2.8%)
100%
a
The recoveries were the average of four spiking levels (0.250, 0.500, 1.00, and 2.00 µg/kg) for all nitrofuran metabolites. Additionally, one spike was analyzed at 2 times the highest calibration standard (10.0 µg/kg) to assay linearity. Quantification was performed by using matrix-matched calibration standards and stable isotopically labeled standards.
b
Recovery data used to calculate the MDL and LOQ for nitrofuran metabolites.
6 VEACH ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 Table 5. Average recoveries and RSDs of chloramphenicol in honey (n = number of replicates)a Fortification level n=7 Analyte CAP
0.300 ng/g 98.6% (6.2%)
b
n=9
n=9
n = 10
n=1
0.500 ng/g
1.00 ng/g
1.50 ng/g
6.00 ng/g
104% (6.2%)
105% (5.5%)
97.7% (5.2%)
111%
a
The recoveries were the average of four spiking levels (0.300, 0.500, 1.00, and 1.50 µg/kg). Additionally, one spike was analyzed at 2 times the highest calibration standard (6.00 µg/kg) to assay linearity. Quantification was performed by using matrix-matched calibration standards and stable isotopically labeled standards.
b
Recovery data used to calculate the MDL and LOQ for chloramphenicol.
Additionally, we could find not published literature for the analysis of FF using an ionization technique other than ESI. Therefore, it was necessary to perform two separate injections on the same extract, using two different LC-MS/MS conditions to achieve the desired MDL and LOQ values. The mass spectrometer was tuned for each targeted compound with respect to response and peak shape. Varying mobile phases and mobile phase compositions were evaluated over the course of this study in an effort to provide optimal performance and sample throughput. Each of the final instrumental acquisition methods developed can be completed within 6 min. Upon analysis completion, the most abundant transition for each analyte of interest was chosen to be used for quantitation. Once the instrument acquisition methods were developed, all of our focus was directed to the sample extraction procedure. Our previous work with honey has shown that honey matrixes differ drastically from each other in nature. This causes varying levels of signal suppression. Therefore, prior to method development, it was deemed necessary to use extracted matrix-matched standards to account for various matrix effects. Furthermore, we used corresponding stable isotopically labeled standards, in combination with matrixmatched standards, for each residue of interest to correct for matrix effects. Our initial efforts in extraction were focused on acid hydrolysis and derivatization through the use of microwave digestion (15). However, although numerous attempts were made, all of our attempts were unsuccessful. This is because the amount of energy required to perform the acid hydrolysis and derivatization process would caramelize the honey samples. When we dropped the energy to a level that would not be detrimental to the sample, it would result in diminished recoveries. Because of the high sugar content of the matrix, it was deemed that traditionally used methods for acid hydrolysis and derivatization should be used (17–22). Varying extraction
solvents were also evaluated. Although no single solvent that was used was optimal for all residues, ethyl acetate was clearly the best universal solvent to achieve adequate recoveries for all our analytes of interest.
Method Validation During method validation, a direct comparison of the time required for the newly developed method with that of the two previously used methods was done. Multiple analysts were involved with the extraction procedure over the validation procedure in order to help demonstrate the method’s ruggedness and to provide an accurate assessment of the amount of time that was reduced using the new method. Over the course of the validation, the average reduction in time was >50%. The validation study was conducted by fortifying three separate honey matrixes at five different concentration levels in accordance with Guidelines for the Validation of Chemical Methods for the FDA Foods Program (24). The three different honey matrixes used during the validation efforts were identified as clover honey, acacia honey, and raw unfiltered honey. The validation occurred over 3 nonconsecutive days and consisted of a total of 36 fortified matrix spikes and 16 blanks. The mean recoveries and RSDs of honey (each NF metabolite, CAP, and FF) are listed in Tables 4–7. Each of the 36 assayed fortified samples met the required confirmation criteria for CAP, FF, and NF metabolites. No falsepositives (0%) were observed in the 16 blanks analyzed for each matrix. This study demonstrates that this procedure is not only capable of increased throughput, but also offers excellent quantitation abilities. This makes it an ideal procedure for laboratories interested in enhanced sample-throughputperforming assays on honey.
Table 6. Average recoveries and RSDs of florfenicol in honey (n = number of replicates)a Fortification level n=7 Analyte FF
4.50 ng/g 94.5% (5.5%)
b
n=9
n=9
n =10
n=1
7.50 ng/g
15.0 ng/g
22.5 ng/g
90.0 ng/g
99.7% (7.5%)
100% (3.0%)
97.9% (4.2%)
105%
a
The recoveries were the average of four spiking levels (4.50, 7.50, 15.0, and 22.5 µg/kg). Additionally, one spike was analyzed at 2 times the highest calibration standard (90.0 µg/kg) to assay linearity. Quantification was performed by using matrix-matched calibration standards and stable isotopically labeled standards.
b
Recovery data used to calculate the MDL and LOQ for nitrofuran metabolites.
VEACH
ET AL.: JOURNAL OF
Table 7. Method detection limits and LOQs Residue
MDL, ng/g
LOQ, ng/g
AHD
0.059
0.189
AMOZ
0.062
0.197
AOZ
0.097
0.310
SC
0.060
0.192
CAP
0.057
0.182
FF
0.739
2.35
Acknowledgments We acknowledge Bryanna Broadaway and Alokita Karmakar (FDA) for their valuable assistance and contributions to the manuscript. References (1) Macherone, A. (2016) Column 18, 2–6 (2) Kasiotis, K.M., Anagnostopoulos, C., Anastasiadou, P., & Machera, K. (2014) Sci. Total Environ. 485–486, 633–642, doi:10.1016/j.scitotenv.2014.03.042 (3) Veach, B.T., Mudalige, T.K., & Rye, P. (2017) Anal. Chem. 89, 3256–3260, doi:10.1021/acs.analchem.6b04889 (4) Ellis, J.D., Evans, J.D., & Pettis, J. (2010) J. Apic. Res. 49, 134–136, doi:10.3896/IBRA.1.49.1.30 (5) Dainat, B., vanEngelsdorp, D., & Neumann, P. (2012) Environ. Microbiol. Rep. 4, 123–125, doi:10.1111/j.17582229.2011.00312.x (6) U.S. Department of Agriculture (2012) United States Honey Production Down 16 Percent, USDA, Washington, DC (7) U.S. Department of Agriculture (2013) National Program 305, Crop Production Systems, Action Plan [2013–2018], USDA, Washington, DC (8) Reybroeck, W., Daeseleire, E., De Brabander, H.F., & Herman, L. (2012) Vet. Microbiol. 158, 1–11, doi:10.1016/j. vetmic.2012.01.012
AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017
(9) Sanford, M.T. (2017) Am. Bee J. XXX (10) U.S. Department of Agriculture (2016) American Foulbrood Disease, USDA, Agricultural Research Service, Beltsville, MD (11) Wilson, M., & Skinner, J. (2013) European Foulbrood: A Bacterial Disease Affecting Honey Bee Brood, eXtension, http:// articles.extension.org/pages/23693/european-foulbrood:-abacterial-disease-affecting-honey-bee-brood (12) Veach, B., Hawk, H., Baker, C., & Linder, S. (2011) FDA/ORA/ DFS Lab. Inf. Bull. 27, 1–15 (13) Sørensen, L.K., Elbæk, T.H., & Hansen, H. (2003) J. AOAC Int. 86, 703–706 (14) Veach, B., Johnson, S., Rimmer, D., Drake, C., Fong, A., Delgado, Q., Broadaway, B., Kibbey, J., & Linder, S. (2014) Inf. Bull. XXX, 1–28 (15) Veach, B.T., Baker, C.A., Kibbey, J.H., Fong, A., Broadaway, B.J., & Drake, C.P. (2015) J. AOAC Int. 98, 588–594, doi:10.5740/jaoacint.14-271 (16) McCalla, D.R. (1983) Environ. Mutagen. 5, 745–765, doi:10.1002/em.2860050512 (17) An, H., Cain, T., Henry, M., Tran, B., Paek, H.C., & Farley, D. (2011) FDA/ORA/DFS Lab. Inf. Bull. 27, 1–35 (18) McCracken, R.J., Spence, D.E., Floyd, S.D., & Kennedy, D.G. (2001) Food Addit. Contam. 18, 954–959, doi:10.1080/ 02652030110050375 (19) Chu, P.-S., & Lopez, M.I. (2005) J. Agric. Food Chem. 53, 8934–8939, doi:10.1021/jf051615o (20) McCracken, R.J., & Kennedy, D.G. (1997) J. Chromatogr. B Biomed. Sci. Appl. 691, 87–94, doi:10.1016/S03784347(96)00448-3 (21) Chu, P.-S., Lopez, M.I., Abraham, A., El Said, K.R., & Plakas, S.M. (2008) J. Agric. Food Chem. 56, 8030–8034, doi:10.1021/jf801398p (22) Lopez, M.I., Pettis, J.S., Smith, I.B., & Chu, P.-S. (2008) J. Agric. Food Chem. 56, 1553–1559, doi:10.1021/jf073236w (23) (2015) European Legislation Regarding Antibiotics in Honey: An Overview, EuroProxima, http://europroxima.com (24) U.S. Food and Drug Administration (2015) Guidelines for the Validation of Chemical Methods for the FDA Foods Program, FDA, Washington, DC (25) U.S. Food and Drug Administration (2003) Fed. Regist. 66, 31938–31939
7
