Content uploaded by Narong Chamkasem
Author content
All content in this area was uploaded by Narong Chamkasem on Aug 02, 2016
Content may be subject to copyright.
FDA/ORA/ORS LIB 4613
Page 1 of 18
U. S. Food and Drug Administration
Laboratory Information Bulletin
Development and independent laboratory validation of an analytical method
for the direct determination of glyphosate, glufosinate, and
aminomethylphosphonic acid in honey by liquid chromatography/tandem
mass spectrometry
Narong Chamkasem, SRL, Atlanta, GA
John D. Vargo, State Hygienic Laboratory at the University of Iowa, Coralville, IA
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/her by appropriate
calibration 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.
ABSTRACT
A simple high-throughput liquid chromatography/tandem mass spectrometry (LC-
MS/MS) method was developed for the determination of glyphosate, glufosinate, and
aminomethylphosphonic acid (AMPA) in honey using a reversed-phase column with weak
anion-exchange and cation-exchange mixed-mode (Acclaim™ Trinity™ Q1). One gram of
sample was shaken with water containing ethylenediaminetetraacetic acid disodium salt
(Na2EDTA) and acetic acid for five minutes. After centrifugation, the supernatant was mixed
with internal standard and directly injected and analyzed in ten minutes by LC-MS/MS with no
sample concentration or derivatization steps. Two precursor/product ion transitions were
monitored in the method for each target compound to achieve true positive identification. Three
internal standards corresponding to each analyte were used to correct for matrix suppression
effects and/or instrument signal drift. The linearity of the detector response was demonstrated in
the range of 2.5 to 250 ng/mL for each analyte with a coefficient of determination (R2) value of
0.998. Through the use of this internal standard calibration method, the average recovery for all
analytes at 25, 50, 100, and 500 ng/g (n = 11) ranged from 87 to 111% with a relative standard
deviation of less than 12%.
INTRODUCTION
Glyphosate (N-phosphonomethyl glycine) and glufosinate [ammonium(S)-2-amino-4-[hydroxyl
(methyl) phosphinoyl] butyrate] are non-selective post emergence herbicides used for the control
of a broad spectrum of grasses and broad-leaf weed species in agricultural and industrial fields.
AMPA is the major metabolite of glyphosate and is classified as a toxicologically significant
compound (1). According to recent reports, there has been a dramatic increase in the usage of
these herbicides, which are of risk to both human health and the environment (2). Glyphosate
and glufosinate have high efficacy, low toxicity, and an affordable price, when compared with
FDA/ORA/ORS LIB 4613
Page 2 of 18
other herbicides. These factors lead to its wide utilization on several crops. Farmers also use
glyphosate as a desiccant to rapidly kill above ground growth of crops such as wheat. This
allows for rapid dry down for easy harvest. The use of glyphosate in agriculture has increased
significantly with the introduction of transgenic crops such as Roundup-Ready® soybeans and
corn, which enable farmers to directly apply low cost broad spectrum herbicide products to their
fields without harming crops. In the United States, glyphosate is currently the most widely used
herbicide, with 180 to 185 million pounds applied in the agricultural sector during 2007, 5 to 8
million pounds used in the home and garden markets, and 13-15 million pounds used in
industrial, commercial and governmental weed control applications (2). This high level of use
has led to concerns about its effects on humans and the environment. A recent study by
researchers from Boston University and Abraxis LLC found significant amounts of glyphosate in
honey (41 out of 69 samples collected in the Philadelphia, US metropolitan area) with a
concentration range between 17 and 163 ng/g using enzyme linked immune sorbent assay
(ELISA) (3). This method is quick, inexpensive, and sensitive; however, it does not have
confirmation method to prevent false positives. A quick, accurate, and sensitive method with a
positive confirmation method to determine these herbicides in honey must be developed to
support regulatory actions.
Glyphosate, glufosinate, and AMPA are very polar compounds and insoluble in organic solvents.
Therefore, the use of classical organic solvent extraction is very difficult and ineffective. As a
result, the isolation and quantification of these herbicides have posed a challenge to the
analytical chemist. Alferness and Iwata used an aqueous extraction method to extract glyphosate
and AMPA from soil, plant and animal matrices (4). This method required the use of lengthy
cleanup procedures that involved both anion and cation exchange columns. Typical silica based
reversed-phase C18 columns experience difficulty with the retention of such polar compounds,
and may generate non-resolved co-eluting peaks, often with polar analytes eluting in the void
volume. The lack of chromophophore or fluorophore also necessitates the use of derivatization
techniques for the determination of these analyte residues by liquid chromatography and gas
chromatography (5-7). Vreeken and co-workers developed an analytical method to analyze
glyphosate, AMPA and glufosinate in water samples using a reversed phase liquid
chromatography separation after pre-column derivatization with 9-fluorenylmethyl
chloroformate (FMOC-Cl) and detection by LC-MS/MS (8). Schreiber and Cabrices streamlined
the derivatization by using a special autosampler for automation to determine these polar
analytes in corn and soybean (9). The derivatization technique is problematic as it requires the
optimization of a number of parameters (temperature, reaction time, concentration and purity of
the reagents, laboratory handling time). Anion exchange, Hydrophilic Interaction Liquid
Chromatography (HILIC), and mixed-mode columns have been used with LC-MS/MS to
determine underivatized glyphosate and other polar pesticides in food matrices with limited
success (10-12). An LC/MS method was developed using a mixed-mode HPLC column
(Acclaim Trinity Q1) to directly determine these three analytes in milk and soybean (13,14). This
method should be applicable for honey as well.
This LIB describes a method validation of an LC-MS/MS method using a negative ion-spray
ionization mode for the direct determination of glyphosate, glufosinate, and AMPA in honey. A
quick and reliable extraction method that requires small sample size, non-toxic solvent, and an
FDA/ORA/ORS LIB 4613
Page 3 of 18
effective sample cleanup procedure to ensure method ruggedness, sensitivity, and selectivity is
provided.
MATERIAL AND METHODS
Chemicals and Materials
Pesticide standards (≥ 99% purity) were purchased from LGC Standards (Manchester, NH)
consisting of glyphosate, AMPA, glufosinate, glyphosate 13C215N (100 µg/mL), AMPA 13C 15N
(100 µg/mL), and glufosinate D3 (1 mg). Acetonitrile and water of HPLC grade were obtained
from Fisher Scientific (Pittsburgh, PA). Formic acid was obtained as 98% solution for mass
spectrometry from Fluka (Buchs, Switzerland.). Acetic acid, ammonium formate and
ethylenediaminetetraacetic acid disodium salt (Na2EDTA) were purchased from Fisher Scientific
(Pittsburgh, PA). Extracting solvent (50 mM acetic acid/10 mM Na2EDTA) was prepared by
mixing 572 µL of acetic acid and 0.74 g of Na2EDTA in 200-mL of purified water. EDP 3
electronic pipettes at different capacities (0-10 µL, 10-100 µL, and 100-1000 µL) were
purchased from Rainin Instrument LLC (Oakland, CA) and were used for standard fortification.
A solution of 500 mM ammonium formate/formic acid (pH 2.9) was prepared as follows: 15.76
g of ammonium formate was dissolved in approximately 300 mL of HPLC water and adjusted
with 98% formic acid (approx. 28.3 mL) until the pH reached 2.9 (using pH meter), and the
solution was diluted to 500 mL with water. The HPLC mobile phase A was HPLC grade water
and mobile phase B was prepared by mixing 100 mL of the 500 mM buffer solution with 900 mL
of purified water (final concentration was 50 mM).
Standard Preparation
The individual stock solutions of glyphosate, glufosinate, and AMPA at 1 mg/mL were prepared
in water. These stock solutions were used to prepare standard mix solutions at 50, 10, 2, 1 and
0.5 ng/uL. The solutions were maintained at 4 °C in polypropylene tubes to avoid adsorption to
glass. The internal standard (IS) solution of glyphosate 13C215N, AMPA 13C15N, and glufosinate
D3 at 2 ng/µL was prepared by dissolving the stock standard in water and stored in a plastic tube.
The mixed standard solutions were further diluted with the extracting solvent to obtain standard
mixes from 2.5 to 250 ng/mL. The calibration standards were prepared in the extracting solvent
with IS solutions for the calibration curves as described in Table 1.
Sample Preparation and Extraction Procedure
Two organic honey samples were obtained from a local market. The samples were weighed at 1
± 0.1 g each in 50-mL plastic centrifuge tubes (Fisher Scientific, Pittsburgh, PA) and fortified
with native standard solutions at 25, 50, 100, and 500 ng/g using Table 2. The samples were
allowed to stand at room temperature for 1 h. Extracting solvent (4.3 mL) was added to each tube
using an automatic pipette. The tubes were capped tightly and shaken for 5 min on a SPEX 2000
Geno grinder (SPEX Sample Prep LLC, Metuchen, NJ) at 2000 stroke/min then centrifuged at
3,000 x rpm for 5 min using a Q-Sep 3000 centrifuge (Restek, Bellefonte, PA). The clear liquid
FDA/ORA/ORS LIB 4613
Page 4 of 18
(180 µL) was pipetted into a plastic 300- µL autosampler vial (Wheaton, Millville, NJ)
containing 20 µL of 2 ng/µL IS solution. The vial was capped and mixed on a vortex mixer for
30 s. A 20 µL of sample was injected into the LC-MS/MS system.
LC-MS/MS Analysis
LC-MS/MS analysis was performed by using a Shimadzu HPLC system. The instrument was
equipped with two LC-20AD pumps, a Sil-20AC autosampler, and a CTO-20AC column oven
(Shimadzu, Kyoto, Japan), coupled with a 6500 Q-TRAP mass spectrometer from AB SCIEX
(Foster City, CA). The Analyst software (version 1.6) was used for instrument control and data
acquisition. Nitrogen and air from TriGas Generator (Parker Hannifin Co., Haverhill, MA) were
used for nebulizer and collision gas in LC-MS/MS. An Acclaim™ Trinity™ Q1 (3 μm, 100 x 3
mm) analytical column from Thermo Scientific (Sunnyvale, CA) and a C18 SecurityGuard
guard column (4 x 3 mm) from Phenomenex (Torrance, CA) were used for HPLC separation at
35 °C with sample injection volume of 20 μL. The mobile phase was 100% A (water) for 30 s at
a flow rate of 0.5 mL/min then ramped up to 100% B (ammonium formate/formic acid buffer)
immediately for 4 min to elute the analytes. The column was equilibrated with 100% A at a flow
rate of 0.7 mL/min for 6 min for a total run time of approximately 10 m. A diverter valve
connected between the HPLC column and the MS interface was used to direct the LC eluent to
waste just before the AMPA peak (2 min) and after the glyphosate peak (3.7 min). The MS
determination was performed in negative electrospray mode with monitoring of the two most
abundant MS/MS (precursor/product) ion transitions using a scheduled MRM program of 60 s
for each analyte. Analyte-specific MS/MS conditions and LC retention times for the analytes are
shown in Table 3. The MS source conditions were as follows: curtain gas (CUR) of 30 psi, ion
spray voltage (ISV) of -4500 volts, collisionally activated dissociation gas (CAD) is high,
nebulizer gas (GS1) of 60 psi, heater gas (GS2) of 60 psi, source temperature (TEM) of 350 ºC.
RESULTS AND DISCUSSION
Optimization of Sample Extraction Procedure
A honey sample (1 g) was spiked with the analytes at 100 ng/g and shaken with 5 mL of water
for 5 min on a Geno grinder at 2000 stroke/min. After the centrifugation, the supernatant was
injected along with standard solution in water at the same concentration using the isocratic
elution with 50 mM ammonium formate buffer solution previously used in the milk and soybean
method (13,14). The result was disappointing due to the poor peak shape and poor response of
AMPA. Honey contains mostly sugars which are polar compounds which tend to coelute with
AMPA near the solvent front. A few sample cleanup procedures were evaluated to eliminate
sugar in the sample. Due to their phosphonate structures, glyphosate, glufosinate, and AMPA
could be retained and efficiently purified on an anion-exchange SPE cartridge while the sugars,
non-ionic compounds, would pass thru. A strong anion solid phase cartridge was previously used
as a cleanup step for the determination of these analytes in beer and tea (15). Three anion
exchange cartridges (NH2, WAX, and SAX 500 mg/6 mL) were evaluated. One gram of honey
spiked with 100 ng of the analytes was shaken with 5 mL of water. One milliliter was loaded on
these cartridges (previously conditioned with water and methanol). The cartridges were washed
FDA/ORA/ORS LIB 4613
Page 5 of 18
with 3 mL of water and eluted with 6 mL of 1:9 1N HCl:methanol. The elutant was analyzed by
LC/MS. Due to the high concentration of methanol in the sample, only 1 μL of the extract was
injected to maintain a good peak shape of AMPA. The result was mixed. Strong anion exchange
SPE (SAX) did not effectively retain AMPA, while amino propyl (NH2) phase lost glufosinate
during the loading step. Weak anion exchange (WAX) SPE gave good overall recovery for all
analytes. However, in order to decrease the detection limit, the sample had to be evaporated to
decrease the sample volume and to evaporate the methanol to improve the peak shape. The
evaporation step was time-consuming and the high concentration of hydrochloric acid left in the
sample may affect the peak shape. Honey solution made of a higher honey concentration (1 g in
3 mL of water) was loaded on the WAX SPE to increase the concentration of the analytes in the
sample. High sugar content reduced effectiveness of the WAX SPE to retain AMPA and resulted
in poor recovery. Because of the poor loading efficiency on the WAX SPE (at high honey
content in the solution) and the long evaporation time (to decrease sample volume), the cleanup
procedure was not further evaluated. The alternative solution to minimize matrix effect was to
modify chromatographic condition to move analyte peaks further away from the solvent front by
using a gradient elution.
Chromatography Optimization
The isocratic elution of the Acclaim Q1 column with 50 mM ammonium formate was previously
developed for glyphosate analysis in milk and soybean (13,14). This condition worked well with
these matrices which did not contain a high concentration of sugar. In order to increase analyte
retention on this column, a much lower salt concentration mobile phase at the initial condition
must be used. Different gradient conditions were evaluated. Finally, the mobile phase condition
was optimized by using a step gradient of 100 % water for 30 sec at a flow rate of 0.5 mL/min to
sufficiently retain AMPA (2.4 min), immediately followed by 100 % 50 mM ammonium formate
(pH 2.9) immediately for another 4 min to elute glufosinate and glyphosate. The column was
then quickly equilibrated with 100% water at a flow rate of 0.7 mL/min for 6 min for a total run
time of 10 min. Honey solution in water (1 g/5 mL) containing the analytes at 250 ng/g was
injected into the LC/MS using the isocratic mode and the step gradient mode for comparison
(Figure 1). The step gradient produced a better peak shape of AMPA, and the analytes eluted
further away from the sample matrix (mostly sugars). This HPLC condition allowed the diluted
honey solution to be directly injected without concentration and achieved good sensitivity and
good peak shape for all analytes. The extracting solvent containing diluted acetic acid and EDTA
provided good extraction of glyphosate in milk by precipitating the protein and preventing
glyphosate from binding with metal ions. It was observed that a sharper peak shape of glyphosate
was obtained when the standard solution contained EDTA instead of water alone. An aqueous
solution of EDTA (50 mM) was previously used to restore the column performance for
glyphosate after a few sets of samples were analyzed on the Acclaim WAX-1. It was believed
that trace metals in the column may broaden the glyphosate peak. EDTA was used to eliminate
the metals ion in the system (10).
The moisture content in honey reported in the literature is approximately 17-20 % (16).
Therefore, 1 g of honey should contain approximately 0.2 mL of water. However, when it was
shaken with 4.8 mL of the extracting solvent, the final volume was 5.5 mL. At least ten different
honey samples showed the same results. This indicated that the extra 0.5 mL may have come
FDA/ORA/ORS LIB 4613
Page 6 of 18
from honey. It was found that in order to adjust the final volume of the solution to 5 mL, only 4.3
mL of the extracting solvent was needed to mix with 1 g of honey. The final volume of 5 mL of
the honey extract was then used for the analyte concentration calculation.
Since there was no sample cleanup step in this method, it was necessary to use a diverter valve to
bypass the HPLC eluate away from the LC/MS interface to waste at the beginning of the run
until just before the AMPA retention time. This step prevented sugar in honey from entering the
heated ion source and forming deposits at the LC/MS interface. The valve also diverted the
HPLC eluate after the glyphosate peak to waste at higher mobile phase flow rate, to keep the
LC/MS interface clean from the sample matrix.
Evaluation of Matrix Effects
Matrix effect (%ME) in the sample extract is calculated as the ratio of the analyte response in the
sample matrix divided by the response of the analyte in the extracting solvent multiplied by 100.
Therefore, a value of 100% means that no matrix effect is present. If the value is less than 100%,
it means that there is matrix suppression. If the value is more than 100%, matrix enhancement
exists. Standard solutions at 50 ng/mL in the extracting solvent and in honey extract (1 g/5 mL)
were prepared and injected to evaluate the matrix effect. Table 4 shows the %ME of all analytes
in honey extract. AMPA and glyphosate demonstrated matrix enhancement approximately 130
and 150%, respectively, while glufosinate had severe suppression (50%). Based on this data,
internal standards are needed for accurate quantification of these analytes.
Method Validation
A method validation at the FDA laboratory in Atlanta, GA was performed using two organic
honey samples collected from a local market. The calibration standard solutions at
concentrations from 2.5 to 250 ng/mL were prepared in the extracting solvent with the addition
of IS (Table 1). These standard solutions were injected along with the fortified samples and
sample blank as described in the Table 2. The accuracy and precision of the method was
evaluated via recovery experiment on two blank honey samples (A and B) spiked at 25, 50, 100,
and 500 ng/g. The specificity of the method was evaluated by analyzing reagent blank, blank
sample, and blank sample spiked at the lowest fortification level (25 ng/g). No relevant signal
(above 30% of the 25 ng/g sample) was observed at any of the transitions selected in the blank
honey sample A. Honey sample B has approximately 10 ng/g of incurred residue of glyphosate.
Glyphosate, a herbicide, is not used to treat bee hives; however, bees may carry glyphosate from
an agricultural area treated with glyphosate. A reagent blank was injected immediately after the
250 ng/mL standard and only glyphosate approximately 0.5 to 0.8% carry-over was observed.
The second injection of the reagent blank showed no trace of glyphosate above 20% of the
lowest standard solution of 2.5 ng/mL. It is advisable to inject a reagent blank after injecting a
sample containing high concentrations of the analytes to minimize false positive for the next
sample.
The sensitivity, expressed in terms of limit of detection (LOD), was estimated as 3 times the
standard deviation of the quantitative MRM response from the replicates fortification of honey
sample at 25 ng/g. The limit of quantification (LOQ) was estimated as 10 times the standard
FDA/ORA/ORS LIB 4613
Page 7 of 18
deviation in a similar manner. Method linearly was determined for each target compounds using
a linear regression curve fit (1/x weighing). The average percent recovery, relative standard
deviation, method LOD/LOQ, and coefficient of determination (R2) are reported for all
compounds in Table 5. Due to the matrix effect (suppression/enhancement), IS must be used for
accurate quantification. Quantification was based on calibration standard in the extracting
solvent with corresponding IS. No matrix-matched calibration standard was required. A
quadratic curve fit (with or without 1/x weighting) is permitted if this provides a better curve fit
for the data.
Accuracy (recovery %) and precision (relative standard deviation or RSD %) for the two honey
samples (A and B) are shown in Table 5. The average recoveries were 87-102% with an RSD of
12% for glyphosate, 90-107% with an RSD of 7.3% for glufosinate, and 90-111% with an RSD
of 6.7% for AMPA. For honey A, the LOD of glyphosate, glufosinate, and AMPA were 5, 5, and
1 ng/g, respectively. The LOQ of glyphosate, glufosinate, and AMPA were 16, 17, and 4 ng/g,
respectively. Since honey B sample has approximately 10 ng/g of incurred residue of glyphosate,
the standard deviation of recovery at 25 ng/g was a bit higher (26 ng/g) than the standard
deviation of recovery found in honey A (16%). Since the honey was very viscous, the
concentration of the incurred residue may not have been constant throughout the sample. The
LOQ of glufosinate in honey B (18 ng/g) was comparable to the LOQ found in honey A. The
LOQ of AMPA in honey B was 16 ng/g as reported in Table 5. Chromatograms of glyphosate,
glufosinate, and AMPA in honey A blank and honey A blank fortified at 25 ng/g are shown in
Figure 2. Chromatograms of a honey sample containing 121 ng/g of glyphosate are shown in
Figure 3. Nineteen honey samples were collected from the local market and a private honey farm
and analyzed by the proposed method (Table 7). Nine samples (47%) contained glyphosate
higher than 16 ng/g (estimated LOQ). Glufosinate and AMPA were not detected in any of the
samples.
Inter-laboratory validation (spiked and incurred samples)
The method was transferred to the second lab for a small scale independent laboratory (State
Hygienic Laboratory at the University of Iowa, Coralville, IA) validation under similar
instrument conditions using a 5500 Q-TRAP from AB-Sciex. It involved the analysis of spiked
samples and samples containing filed-incurred residues. This is to prove if the method
performance could be duplicated at a different laboratory. Accuracy and reproducibility data
were obtained by replicate analysis of spiked sample (blank honey spiked at 25, 50, and 100
ng/g, n = 7,4, and 4). The results are presented in Table 6. The mean recovery for all analytes
ranged from 84 to 108% with the RSD ranged from 2.3 to 12.8%. The extraction procedure and
HPLC conditions used at the state lab were the same as the FDA lab with no modification. Three
honey samples (honey from FL, LA, and IA) containing incurred residue of glyphosate were
analyzed by both laboratories and the results were comparable, indicating that the method was
reliable for use (Table 7).
CONCLUSION
This work describes a five-minute extraction with aqueous solution of acetic acid and Na2EDTA
which provides for rapid and direct determination of glyphosate, glufosinate, and AMPA residue
FDA/ORA/ORS LIB 4613
Page 8 of 18
in honey samples. After the centrifugation to obtain a clear extract, the sample was injected
directly to the LC/MS/MS system. The mixed-mode Acclaim™ Trinity™ Q1 HPLC column
allows the analytes to be retained on the column and separated from each other without a
derivatization step. The step gradient elution developed in this method improved the peak shape
and retention of the analytes over isocratic elution. Sugar, the major component in the honey
sample, was eluted much earlier and diverted to waste to prevent severe matrix suppression of
AMPA and keep the ion-source clean. Negative mode ion-spray with MS/MS measurement
gives excellent sensitivity and selectivity that produce distinct chromatographic peaks with
minimal interference. The use of internal standard for each analyte minimized the matrix effect
and provides accurate quantification. The in-house and the inter-laboratory validation studies,
using spiked blank honey and honey with incurred residue of glyphosate, demonstrated that the
method is quick, rugged, selective, and sensitive enough to determine glyphosate, glufosinate
and AMPA in honey at or above the 25 ng/g level. It can be used as an alternative method to the
ELISA technique as well as to the traditional FMOC derivatization methods which are tedious
and time-consuming.
FDA/ORA/ORS LIB 4613
Page 9 of 18
LITERATURE CITED
1) EUROPEAN COMMISSION HEALTH & CONSUMER PROTECTION
DIRECTORATEGENERAL. Review report for the active substance glyphosate.
http://ec.europa.eu/food/plant/protection/evaluation/existactive/list1_glyphosate_en.pdf
(accessed Jan 27, 2015)
2) United States EPA 2007 Pesticide Market Estimates
http://www.epa.gov/opp00001/pestsales/07pestsales/table_of_contents2007.htm
(accessed Jan 1, 2015)
3) Robio, F.; Guo, E., Kamp, L. Survey of glyphosate residues in honey, corn and soy
products. J. Envion Analytical Toxicol 2014 4:249.
4) Alferness, P.L.; Iwata Y. Determination of glyphosate and (aminomehtyl) phosphonic
acid in soil, plant and animal matrices, and water by capillary gas chromatography with
mass selective detection. J. Agric. Food Chem. 1994, 42, 2751-2759
5) Qian, K.; Tang, T.; Shi, T.; Li, P.; Cao, P. Solid-phase extraction and residue
determination of glyphosate in apple by ion-pairing reverse-phase chromatography with
pre-column derivatization . J. Sep Sci. 2009, 32(4), 2494-2300
6) Ibanez, M, P; Ozo, O.J.; Sancho, J.V.; Lopez, F.J.; Hernandez, F. Residue determination
of glyphosate, glufosinate and aminomethylphosphonic acid in water and soil samples by
liquid chromatography coupled to electrospray tandem mass spectrometry . J.
Chromatogr A. 2005, 1081 (2), 145-155.
7) http://www.crl-pesticides.eu/library/docs/srm/meth_QuPPe.pdf (accessed Jan 27, 2015)
8) Vreeken, R. J.; Speksnijder, P.; Bobeldijk-Pastorova, I. & Noij, Th. H. M. (1998).
Selective analysis of the herbicides glyphosate and amonimethylphosphonic acid in water
by on-line solid-phase extraction-high-performance liquid chromatography electrospray
ionization mass spectrometry. 1988 Journal of Chromatography A, 794, 1-2,
187-199,
9) Schreiber, A.; Cabrices, O.G. Automated derivatization, SPE cleanup and LC-MS/MS
determination of glyphosate and other polar pesticides : application note 8013813-01
http://www.newfoodmagazine.com/wp-content/uploads/Glyphosate_QTRAP-
4500_Gerstel_AB-SCIEX_8013813-01.pdf
10) Hao, C.; Morse, D.; Morra, F.; Zhao, X.; Yang, P.; Nunn B. Direct aqueous
determination of glyphosate and related compounds by liquid chromatography/tandem
mass spectrometer using reversed-phase and weak anion-exchange mixed-mode column.
J. Chromatogr A. 2011, 1218 (33), 5638-5643.
FDA/ORA/ORS LIB 4613
Page 10 of 18
11) http://www.crl-pesticides.eu/library/docs/srm/meth_QuPPe.pdf (accessed Jan 27, 2015)
12) Botero-Coy, A.M.; Ibanez, M.; Sancho, J.V.; Hernandez, F. Direct liquid
chromatography–tandem mass spectrometry determination of underivatized glyphosate in
rice, maize and soybean. J. Chromatogr A. 2013, 1313 (1), 157-165
13) Chamkasem, N.; Harmon, T.; Morris, C. Direct determination of glyphosate, glufosinate,
and AMPA in milk by liquid chromatography/tandem mass spectrometer. Laboratory
Information Bulletin U.S. Food and Drug Administration, Office of Regulatory Affairs,
Rockville, MD (LIB 4595).
14) Chamkasem, N.; Harmon, T.; Morris, C. Direct determination of glyphosate, glufosinate,
and AMPA in milk by liquid chromatography/tandem mass spectrometer. Laboratory
Information Bulletin U.S. Food and Drug Administration, Office of Regulatory Affairs,
Rockville, MD (LIB 4596).
15) Nagatomi, Y.; Yoshioka, T.; Yanagisawa, M.; Uyama, A.; Mochizuki, N. Simultaneous
LC-MS/MS analysis of glyphosate, glufosinate, and their metabolic products in beer,
barley tea, and their ingredients. Biosci. Biotechnol. Biochem. 2013 vol 77(11) 2218-
2221.
16) https://en.wikipedia.org/wiki/Honey#Nutritional_and_sugar_profile (access Jan 5, 2016)
FDA/ORA/ORS LIB 4613
Page 11 of 18
Table 1. Preparation of calibration standard solutions.
Standard in extracting solvent (ng/mL) 2.5 5 10 50 100 250
Volume of standard solution in vial (µL) 180 180 180 180 180 180
IS 2 ng/µL (µL) 20 20 20 20 20 20
Total volume (µL) 200 200 200 200 200 200
IS concentration (ng/mL) 100 100 100 100 100 100
Standard concentration equivalent (ng/mL) 2.5 5 10 50 100 250
Table 2. Preparation of fortified samples (for each 1 g of sample and a final volume of
5 mL).
fortification level standard mix standard mix expected conc.
(ng/g) (ng/uL) used (µL) in the extract (ng/mL)
0 0 50 0
25 0.5 50 5
50 1 50 10
100 2 50 20
500 10 50 100
FDA/ORA/ORS LIB 4613
Page 12 of 18
Table 3. Retention time and MRM conditions for LC-MS/MS analysis a.
Analyte Precursor Product DP CE EP CXP Retention
Ion (m/z) Ion (m/z) Time (min)
AMPA.1 110 63 -60 -24 -10 -10 2.4
AMPA.2 110 79 -60 -26 -10 -10 2.4
AMPA 13C15N (IS) 112 63 -60 -24 -10 -10 2.4
Glufosinate.1 180 95 -46 -23 -10 -10 2.6
Glufosinate.2 180 85 -46 -26 -10 -10 2.6
Glufosinate D3 (IS) 183 63 -46 -26 -10 -10 2.6
Glyphosate.1 168.2 63 -110 -30 -10 -10 2.8
Glyphosate.2 168.2 79 -110 -55 -10 -10 2.8
Glyphosate 13C215N (IS) 171 63 -110 -30
-10 -10 2.8
a. Compound dependent parameters: DP = declustering potential, CE = collision
energy, EP = entrance potential, CXP = collision cell exit potential
Table 4. Matrix effect evaluation (peak area of standard in matrix vs. in solvent).
AMPA (area) Glufosinate (area) Glyphosate (area
native IS native IS native IS
solvent spike (50 ng/mL) 21100 24200 39000 20200 22800 87000
organic honey spike (50 ng/mL) 27400 31900 20100 10000 34100 130000
matrix effect (%) 130 132 52 50 150 149
FDA/ORA/ORS LIB 4613
Page 13 of 18
Table 5. Recovery (%) and RSD (%) data obtained in the single laboratory validation
experiments (FDA laboratory in GA).
Analyte Fortification Recovery RSD LOD LOQ Linearity
level (ng/g) (%) (%) ng/g ng/g R square
Honey A, honey from Ivory Coast, (n = 4)
Glyphosate 25 92 7.0 5 16 0.9997
50 102 2.1
100 92 2.6
500 96 2.8
Glufosinate 25 107 6.3 5 17 0.9991
50 107 5.3
100 91 3.0
500 98 2.7
AMPA 25 106 1.6 1 4 0.9998
50 104 3.0
100 90 3.4
500 103 3.1
Honey B a , organic honey. (n =7)
Glyphosate 25 90 12.1 8 26 0.9981
50 93 4.6
100 87 8.0
500 102 8.1
Glufosinate 25 107 6.8 5 18 0.9990
50 94 5.9
100 90 8.2
500 101 7.3
AMPA 25 111 5.6 5 16 0.9994
50 103 3.5
100 93 6.7
500 103 3.5
a Honey B contain approximately 10 ng/g of glyphosate residue.
FDA/ORA/ORS LIB 4613
Page 14 of 18
Table 6. Recovery (%) and RSD (%) data obtained in the second laboratory validation
(State Hygienic Laboratory at the University of Iowa).
Spike Level (ng/g) n Recovery (%)
Range
Average
Recovery (%) RSD (%)
Glyphosate 25 7 73 - 105 84 12.8
50 4 79 - 89 86 4.9
100 4 90 - 97 95 3.2
Glufosinate 25 7 88 – 106 97 6.6
50 4 91 – 107 98 8.1
100 4 91 – 112 103 9.0
AMPA 25 7 104 – 112 108 3.9
50 4 101 – 107 104 2.8
100 4 103 – 108 105 2.3
FDA/ORA/ORS LIB 4613
Page 15 of 18
Table 7 Glyphosate found in honey samples collected from the local market.
Sample FDA lab a State Lab IA b Source
glyphosate (ng/g) glyphosate (ng/g)
wild flower honey < 16 GA, USA
organic honey 17 Brazil
orange blossom honey < 16 FL, USA
clover honey 26 GA, USA
orange blossom honey 21 NC, USA
clover honey 40 GA, USA
clover honey < 16 Canada
wild flower honey/miel 46 Canada
Gourmet honey < 16 unknown
Manuka honey < 16 New Zealand
honey < 16 France
honey from Canada 19 Canada
Unknown honey < 16 Ivory Coast
honey blend with fructose/flavor < 16 Taiwan
local honey < 16 GA, USA
organic honey < 16 TX, USA
honey from FL 24 24 FL, USA
honey from LA 121 123 LA, USA
honey from IA 35 42 IA, USA
353293 653 IA, USA
353294 34 IA, USA
353295 85 MN, USA
353296 103 IA, USA
353297 23 IA, USA
353298 40 IA, USA
IA-1 23 IA, USA
IA-2 < 10 IA, USA
IA-3 44 IA, USA
a LOQ = 16 ng/g
b LOQ = 10 ng/g
FDA/ORA/ORS LIB 4613
Page 16 of 18
Figure 1 Chromatograms of honey samples (spiked at 250 ng/g) analyzed by isocratic
and step gradient elution. (peak 1 = AMPA, peak 2 = glufosinate, peak 3 = glyphosate, peak
4 = honey matrix)
FDA/ORA/ORS LIB 4613
Page 17 of 18
Figure 2 Chromatograms of honey (A) blank (left) and honey blank + 25 ng/g (right).
Glyphosate channel
Glufosinate channel
AMPA channel
FDA/ORA/ORS LIB 4613
Page 18 of 18
Figure 3 Chromatograms of a honey sample containing 121 ng/g of glyphosate.
Glyphosate channel
Glufosinate channel
AMPA channel
