Confirmation of Multiple Phenicol Residues in Honey by Electrospray LC/MS
Volume 18, No. 5, May 2002
Sherri Turnipseed, Cathy Burns, Joseph Storey, Rebecca Lee, Al Pfenning
Animal Drugs Research Center and Denver District Laboratory
Food and Drug Administration, P.O. Box 25087, Denver CO 80225-0087
The Laboratory Information Bulletin is a tool for the rapid dissemination of laboratory methods (or information) which appear to work. It may not report completed scientific work. The user must assure him/herself by appropriate validation procedures that LIB methods and techniques are reliable and accurate for his/her intended use. Reference to any commercial materials, equipment, or process does not in any way constitute approval, endorsement, or recommendation by the Food and Drug Administration.
The analysis of honey for chloramphenicol and related compounds is important for several reasons. These drugs may be found in honey because during honey production chloramphenicol may be applied to hives to reduce the incidence of diseased colonies (1). Residues of chloramphenicol (CAP) are of particular concern because this drug can cause serious acute reactions, including aplastic anemia, in susceptible individuals (2). Recently is has been reported that chloramphenicol has been found in several foodstuffs from Asia, including honey (3).
There are limited reports of the analysis of CAP and other phenicols in honey or other substances using electrospray LC/MS (4). Several others government (EU, Canadian, state government) methods have also been reported, but are not published in the open literature. Our laboratory has been working with these compounds for many years. The traditional approach to the determination and confirmation of these compounds is isolation from tissue or fluids using liquid/liquid extraction, derivatization with silylating agents to form volatile derivatives, and analysis by GC/ECD and/or GC/MS with negative chemical ion detection (5-7). Recently we have worked on a confirmatory method for these drugs in shrimp using LC/MS with electrospray (8).
The scope of this method is to describe a confirmatory (qualitative) method for chloramphenicol (CAP) and several related compounds (florfenicol and thiamphenicol) in honey using negative ion electrospray with ion trap LC/MSn analysis. Because the chromatographic and MS conditions were initially developed to look for the metabolite florfenicol amine as well as these other drugs, the method allows for detection of this compound in the first part of the chromatographic run, but the presence of the amine was not specifically evaluated for in honey.
- Roybal, J.E. "Chloramphenicol and Related Drugs" in Analytical Procedures for Drug Residues in Food of Animal Origin (1998) ed, S.B. Turnipseed and A.R. Long, Science Technology System, W. Sacramento, CA pp. 227-260. (primary reference?)
- Hormazabal, Y J. Liq Chromatogr. & Related Technique
- Pfenning, A.P., Roybal, J.E., Rupp, H.S., Turnipseed, S.B., Gonzales, S.A., Hurlbut, J.A. (2000) J.AOAC Int. 83, 26.
- Pfenning, A.P., Madson, M.R., Roybal, J.E, Turnipseed, S.B., Gonzales, S.A., Hurlbut, J.A., Salmon, G.D (1998) J. AOAC Int. 81, 714.
- Kijak, P.J (1994) J.AOAC. Int. 77, 34.
- Turnipseed, S.B., Roybal. J.E., Pfenning, A.P., Kijak, P.J. in preparation for Anal Chim Acta
- Extraction Ten grams of honey is dissolved in 10 mL of water, extracted with 20 mL ethyl acetate, shaken and centrifuged. The extraction steps are repeated and the ethyl acetate layers are evaporated to dryness. Thirty mL water is added to the flask, followed by shaking, centrifugation and two hexane rinses. The aqueous phase is passed through a series of SPE columns. The analyte is extracted off the final SPE with methanol. The methanol is evaporated to dryness. The extracts are reconstituted into a small volume of 0.1% formic acid and filtered into LC vials.
- Mass Spectral Analysis
A. Qualitative Confirmation
The qualitative confirmation of phenicols in honey is based on unique mass spectral characteristics of these compounds as evaluated by established guidelines (9,10). One unique aspect of these compounds is the fact that they contain two chlorine atoms, thus giving rise to unique isotopic patterns. In order to take advantage of this fact, the MS2 spectra is obtained not only from the parent ion ([M-H]-, but also from the corresponding M+2 (35Cl37Cl) isotope peak. For example, in the MS2 spectra of CAP ([M-H]- pair m/z 321/323) the predominant ion is m/z 194 which corresponds to [M- H-(NH2COCCl2H)]-. Also present in this spectra are the ions m/z 176 [m/z 194 - (H2O)]- (15%) 249 (30%), [M-H-(2HCl)]-, and 257 [M-H-(HCOCl)]- (25%). These ions are also present in the MS2 spectra of m/z 323, although the peak at 257 is split (into peaks of approximately equal abundance) between ions at m/z 257 and 259, indicating the loss of one chlorine atom (either 35Cl or 37Cl) from the 35Cl37Cl parent ion.
The florfenicol MS2 spectra is dominated by the loss of HF from the parent ions. This is observed as m/z 335.8 when the 35Cl35Cl parent ion (m/z 356) is isolated or m/z 337.8 when the 35Cl37Cl ion is fragmented. To obtain additional confirmatory ions MS3 is performed on ion 335.8 to give a spectra which includes the ions 219 (usually 100%, ), as well as m/z 119, 184, 264. Thiamphenicol [M-H]- equal to 354/356 fragments to give the following ions m/z 227, 240, 270, and 290/292.
Confirmation of the florfenicol amine was investigated in the original confirmation method for phenicols in shrimp, and although some steps in this procedure (i.e. use of PRS column, initial 5 min of LC/MS program)are in place to monitor for this drug, the analysis of florfenicol amine was not evaluated in honey.
- Solvents: Ethyl acetate , hexane, acetonitrile (high purity for UV, LC, Burdick & Jackson or equivalent).
- Formic acid used to prepare the mobile phase was purchased from Baker (88%).
- Solid-phase extraction columns: C18: Varian Mega-BE C18, 1g, 6 mL
PRS: Varian Bond-Elut LRC-PRS 500mg
- Syringe filters: 4 mm syringe filter 0.45 µm, PFTE. Phenomenex P/N AFO-0422
- Ion Trap LC/MS: The instrument used was a Finnegan LCQ DECA Ion Trap Mass Spectrometer coupled to a modular Spectrasystem LC system. The components of the LC system include a SCM1000 degasser, P4000 LC pump, AS3000 autosampler, and a UV6000LP UV/VIS detector. The software used was Xcaliber Version 1.2.
- LC Column: The LC Column was an Xterra phenyl (2.1 × 100 mm, 3.5 µ, Waters Corp. P/N 186001180). Other phenyl columns would also be acceptable. In this laboratory an Inertsil phenyl (2 × 150 mm, 5 µ, Phenomenex Corp. P/N 0301-150X020) was also tested during method development. If other columns are used, the time segments in acquisition program need to be adjusted to account for shift in retention times.
- Other: Rotoevaporator: with ice trap and water bath set at 50°C
Nitrogen evaporator: 12-sample nitrogen evaporator, with 50°C water bath
Plasticware: 15 mL disposable, conical polypropylene with screw cap
Glassware: pear shape flask, Pasteur pipettes
- Standard Preparation
The compounds were purchased or obtained from: Chloramphencol (USP), Florfenicol , Thiamphenicol (Sigma), Florfenicol and Florfenicol Amine (Schering-Plough).
Fortification Standards: For fortification of honey, individual stock solutions of drug at 1000 µg/mL (1000 ng/µL) were made up in acetonitrile. A combined intermediate standard solution (10 ng/µL) was made by pipetting 1 mL of each individual stock solution into 100 mL volumetric flask and diluting to volume with acetonitrile. Prepare fortification standards, as applicable: Pipet 0.5, 0.2, or 0.1 mL combined standard solution into 10 mL volumetric and dilute to volume with acetonitrile for 5, 2, and 1 ppb fortification standards, respectively.
MS Standards: For MS analysis, stock solutions of drug at 100 µg/mL (100 ng/µL) were made up in methanol. A mixed intermediate standard (1 ng/µL of each drug) was made by diluting 500 µL of each stock solution to 50 mL with 0.1% formic acid.
Working LC/MS Standards. As applicable, LC/MS standards were made as follows:
µL of 0.1%
[ng/µL] equivalent in
1000 4000 0.2 5 400 4600 0.08 2 200 4800 0.04 1 100 4900 0.02 0.5* Assuming 10 g of honey is processed and final extract volume is 250 µL.
Stability: Working LC/MS standard are stable for at least one week.
- Sample Preparation
Control Samples: At least one control (matrix blank) sample should be run with every set of samples.
Fortified Samples: At least two fortified sample should be run with every set of incurred or unknown samples. The concentration of the fortified sample should be in the range of 1-5 ppb.
Incurred Samples: Were not available during method development.
- Sample Extraction
Weigh out 10.00 g (±0.05 g) honey in a 59 mL polypropylene centrifuge tube. Add 10 mL of water, vortex and/or shake until honey is dissolved. Add 20 mL ethyl acetate, shake vigorously for one minute Centrifuge at 4000 rpm for 5 minutes.
Transfer ethyl acetate layer into a 100 mL pear-shaped (P-S) flask. Repeat with a second 20 mL aliquot.
Roto-evaporate at 50° C to dryness. Add 30 mL water to dried flask and vortex for 5 seconds followed with sonication for 2 minutes. Transfer to 30 mL polypropylene centrifuge tube. Rinse P-S flask two times with 5 mL hexane and add to centrifuge tube.
Shake centrifuge tube vigorously for 30 seconds, centrifuge at 4000 RPM for 3 minutes. Remove and discard hexane layer. Repeat with a second 5 mL aliquot.
Apply aqueous phase to conditioned SPE system (PRS column on top of C18 column with optional 75 mL reservoir on top) and allow to drain at 2 drop/second. Discard all liquid. When the level reaches the top of the PRS column bed, add 1 mL water. When the level again reaches the top of the PRS column, discard the column. When the level reached the top of the C18 column, add 2 mL water. Allow to drain and discard the liquid. Dry column under vacuum for 1 minute.
Elute column with 4 mL methanol into a 15 mL polypropylene centrifuge tube. Take to dryness.
The dried extracts are reconstituted into 250 µL of 0.1% formic acid, and filtered for injection into LC-MS system.
- Instrument Operating Parameters
Regardless of the instrument used, certain performance verification criteria should be incorporated into the operating parameters. These include mass calibration, tuning, and appropriate fragmentation patterns. Mass axis calibration should be performed according to the instrument manufacturers' specifications or according to internal laboratory MS standard operating procedures. Signal optimization (tuning) should be adjusted to maximize the abundance of ions of interest . Daily system suitability requirements (described in section F7) should also be met. The following describes the specific operating procedures for the instrument used to validate this method in the developer's laboratory.
A printout of the Instrumental Method Program used is included in Appendix 1.
- Instrumental Configuration: LC/MS analysis is performed using a LCQ DECA mass spectrometer coupled to a TSP P4000 LC via an electrospray interface. The instrument is operated using positive and negative ion detection. The instrument was calibrated according to the manufacturer's instructions. The response for CAP was optimized by tuning on ion m/z 321. For tuning, CAP (1 ng/µL in mobile phase) was pumped through a syringe pump at 10 µL/min and then introduced into the LC flow (250 µL/min 80/20 0.1% formic acid/acetonitrile) via a T before entering the MS source. A representative tune report, along with typical electrospray voltages and settings, is shown in Appendix 2. In the tune file the MS parameters were set to a prescan of 2 and a maximum inject time of 100 ms. The MS2 parameters were also optimized using the tune function of the instrument. For this mode the prescan was set to 1 with a maximum inject time of 500 ms . The collision energy was optimized for both total MS2 ion current, as well as for specific ions (m/z 194, 249) with no significant differences (optimal collision energy was 24-26% in all cases).
- Monitored Response: Using the ion trap, MS2 was performed on the molecular ions for each of the analytes according to the following program:
Program 1: Fixed MS2 Acquisition
Isolation width was set to 2 amu for all MS2 transitions. Positive ion tune should be used for time segment 1 if used. Tune file developed for CAP (described above) should be used for other time segments.
Time Segment 1: 2-5 minutes FFA (CAN DELETE THIS SEGMENT)
Scan Event 1: (+) MS [m/z 180-350]
Scan Event 2: (+) MS2 of m/z 248.1 (24% CE) [m/z 65-250]
Scan Event 3: (+)MS3 of m/z 248.1 (24%CE) → m/z 230.1 (32% CE) [m/z 60-250]
Time Segment 2: 5-11 minutes TAP
Scan Event 1: (-) MS [m/z 320-375]
Scan Event 2: (-) MS2 m/z 354.2, (CE 35%) [m/z 65-250]
Scan Event 3: (-)MS2 356.2 (CE 35%)
Time Segment 3: 11-12.5 minutes FF
Scan Event 1: (-) MS m/z 320-375
Scan Event 2: (-) MS2 m/z 356.2,(CE 24%)
Scan Event 3: (-)MS2 m/z 358.2 (CE 24%)
Scan Event 4: (-) MS3 of m/z 356.2 (24%CE) → m/z 335.8 (20% CE)
Time Segment 4: 12.5-18 minutes CAP
Scan Event 1: (-) MS m/z 300-350
Scan Event 2: (-) MS2 m/z 321.2 (CE 24%)
Scan Event 3: (-)MS2 m/z 323.2 (CE 24%)
A UV/Vis diode array detector was also utilized with a scan range of 190-800 nm and channel A set to 270 nm (bandwidth 9 nm) and channel B set to 236 nm (bandwidth 9 nm).
- Specific Operating Conditions: The electrospray interface was operated with a temperature of 275°C. The sheath gas was nitrogen at approximately 35 psi; the auxiliary gas was also nitrogen at approximately 6 psi (optimized for CAP signal). The mobile phase was at flow of 250 µL/min and a column oven was not used. Automated injections of 75 µL were made using "push loop" type injection. The LC flow was diverted away from the mass spectrometer for the first minute. The MS was on from 1-18 minutes. The chromatographic gradient is as follows:
0-5 * 2 98 6-18 20 80 20-22 90 10 23-28 2 98
*note - if not interested in florfenicol amine, chromatographic program could begin at 20% acetonitrile. Time windows might need to be adjusted.
- Procedures for Instrumental Analysis of Samples, Controls, and Standards
Standards are to be run with each set of samples (at the beginning and end of a set of samples, and in the middle of the sequence if many samples are being analyzed). At least two positive controls, i.e. fortified matrix should be run along with any unknown sample extracts. A blank matrix sample (negative control) should also be run along with any unknown sample extracts and must demonstrate the absence of CAP. At least one of the fortified matrix control samples must demonstrate the all confirmation criteria in Section G2v. A solvent blank (mobile phase) should be run before each sample to ensure that there was no carryover from the previous sample or standard. Solvent blanks are not required between duplicates of the same test sample, or when a fortified sample of higher concentration than a previous fortified sample is analyzed.
For qualitative analysis, the important factor is to obtain information to determine if the data meet the confirmation criteria described in Section G2v. Examples of data that can be extracted for this purpose are included in Appendix 3. Ion chromatograms from the full MS (m/z corresponding to [M-H]-) and from MS2 (m/z m/z 194 corresponding to [M-H-(NH2COCCl2H)]- from both fragmentation of both m/z 321 and 323) can be shown along with the the MS2 spectra averaged across the chromatographic peaks. In addition, extracted ion chromatograms for several ions (m/z 194, 257/259, 249, and 176) in the MS2 spectra of 321 and 323 can be shown. As scan data are obtained, relative abundances of representative ions can be estimated from the appearance of the MS2 spectra, or from tabulation data. Integration of ion chromatograms is not necessary.
- System Suitability
The instrument should meet calibration and tuning criteria as described above.
In addition, for each day's analysis, a standard mixture should be analyzed initially to determine the performance qualifications, or system suitability of the instrument. The analytes need to elute at the correct retention time; within ± 5% of what was observed for standards previously (unless column or mobile phase have been changed) and within the time-dependent window if used. It may require one or two injections of standard for compounds to elute at correct retention time if instrument has not been used recently. In addition, the response for 75 µL injection of a 1 ppb standard for CAP should be > 200,000 counts for the 321 ==> 194 MS2 transition.
- Validation Data: Validation data for ion trap MS confirmation of multi phenicol residues in honey are shown in Appendix 3.
- Parameters Evaluated
- Recovery: Fortified samples were analyzed in the range from 2 to 5 ppb with recoveries from 46 percent to 66 percent.
- Reproducibility: A series of standard injections (75 µL injection size) were analyzed using the following standards: At 1 ppb (3 ng on-column) the reproducibility of standard injections as measured by the CAP 321 to 194 transition was 16% (n=6), at 0.25 ng (750 pg on-column), 19.9% (n=5) and at 0.1 ppb (300 pg on-column) it was 40.0% (n=4).
- Specificity: This method meets the specificity guidelines for confirmation methods outlined by Sphon (9) and recently elaborated in CVM's draft guidance (10). During the course of this investigation, several lots of control honey was analyzed and there were no significant interfering peaks in any of the control tissue samples analyzed using the mass filters as described.
- J.A. Sphon, J. Assoc. Off. Anal. Chem. 61, 1247 (1978)
- Center for Veterinary Medicine (2001) Guidance for Industry: Mass Spectrometry for Confirmation of the Identity of Animal Drug Residues
- Sensitivity: For CAP, the ion trap instrument was able to confirm approximately 300-500 pg of standards on-column and honey tissue fortified at 1 ppb was confirmed with a 75 µL injection volume (final extract volume of 250 µL).
- Accuracy, Proof of Recovery from Authentic Samples
Using an ion trap instrument the following criteria must be met for positive qualitative confirmation: For chloramphenicol: 1) The ion m/z 194 [M-H-(NH2COCCl2H)]- must be observed in the MS2 spectra from both parent ions (m/z 321 and 323), and should be a predominant peak in the mass range m/z 100-270. 2)In addition, at least one of the other structurally significant lower abundance ions (m/z 257/259 [M-H-(HCOCl)]-, m/z 249 [M-H-(2HCl)], or m/z 176 [m/z 194- (H2O)]-) must also be present in at least one of the MS2 spectra at an approximate relative abundance to the base peak m/z 194 as is observed in the external standards, and 3) the retention time should be ±5% of external standards run on that day.
The qualitative criteria for the other phenicols is similar. The florfenicol MS2 spectra is dominated by the loss of HF from the parent ions to give only one ion (335.8 from m/z 356 or m/z 337.8 from m/z 358). To obtain additional confirmatory ions MS3 is performed on m/z 337.8. For thiamphenicol [M-H]- (354/356) fragments to give several ions m/z 227, 240, 270, and 290/292. At least two of these should be observed in MS2 spectra from each parent isotope peak. In addition, the retention times for these other residues must also be ± 5% of what is observed from external standards analyzed on the same day.
- Practicality, Sample Throughput, Solvents and Time Requirements: Extraction and LC/MS analysis of 6-8 samples can be accomplished in one day/overnight. For example, initial extraction can be performed in 5 hours. Each LC/MS run takes 28 minutes therefore 6 sample analyses (bracketed by analysis of standards, separated by solvent blanks) can be done in 8-12 hours.
Quality Control Points
- Critical Points
- Extraction: When filtering, be careful that the syringe filter does not disengage.
- Chromatography: A formic acid/acetonitrile mobile phase at 0.25 mL/min on a semi-micro phenyl column resulted in the best chromatographic performance and electrospray sensitivity. The migration of peaks, especially at the beginning of the chromatographic analysis, can be a problem and several injections of standard may be necessary to allow compounds to "settle" into reproducible retention time. Retention times are stable during continuous sequences, even as long as 40-50 samples.
- Mass spectral analysis: In addition to obtaining good agreement between samples and standards analyzed on the same day, a review of the data shows that the relative abundances of ions obtained different days is also very reproducible.
- Performance Specifications: Performance Specifications are outlined above in sections F.4.ii (tuning of mass spectrometer), F.7 (system suitability for standards) and G.2.v (criteria for confirmation)
- Stability: Stability of residues in honey stored for extended periods of time was not evaluated.
- Safety: Standard laboratory safety practices (lab coats, eye protection) should be followed. In addition any safety precautions listed in the determinative SOP for preparation of reagents should be followed. Also follow instrument manufacturers guidelines for safe operation of electrospray LC/MS (particularly with respect to high voltages, high current, and high temperatures).