Identification of fonofos metabolites in Latuca sativa, Beta vulgaris, and Triticum aestivum by packed capillary flow fast atom bombardment tandem mass spectrometry.

Article abstract of DOI:10.1021/jf011052q

The metabolism of fonofos, a thiophosphonate insecticide, was investigated in mature lettuce (Latuca sativa), beet (Beta vulgaris), and wheat (Triticum aestivum). Six new metabolites were identified by LC-MS and LC-MS-MS analysis using fast atom bombardment (FAB) and packed capillary LC columns with application of the on-column focusing technique. These methods provided the sensitivity required to identify unknown metabolites that were present in the mature plants at only 20-230 ppb. Structural elucidation was facilitated by use of fonofos labeled with both carbon-14 and carbon-13 in the phenyl ring. In all three plants fonofos was converted to a glucose conjugate of thiophenoxylactic acid. Oxidation of the glucose conjugate gave isomeric sulfoxides in all species examined. Thiophenoxylactic acid was found esterified to malonic acid in lettuce. In beets, S-phenylcysteine was found as its malonic acid amide. A second metabolite unique to beets was N-(malonyl)-[2[(ethoxyethylphosphinothionyl)oxy]phenyl]cysteine. This novel structure was confirmed by synthesis.

Full text of DOI:10.1021/jf011052q

1922 J. Agric. Food Chem. 2002, 50, 1922−1928
Identification of Fonofos Metabolites in Latuca sativa, Beta
vulgaris, and Triticum aestivum by Packed Capillary Flow Fast
Atom Bombardment Tandem Mass Spectrometry
BRUCE C. ONISKO,*,† DOUG R. TAMBLING,‡ GREG W. GORDER,§
DAVID G. DIAZ,# JOHN L. ERICSON, MIKE P. PRISBYLLA,| AND
X
CHUCK J. SPILLNER
Syngenta, 1200 South 47th Street, Box 4023, Richmond, California 94804-0023
The metabolism of fonofos, a thiophosphonate insecticide, was investigated in mature lettuce (Latuca
sativa), beet (Beta vulgaris), and wheat (Triticum aestivum). Six new metabolites were identified by
LC-MS and LC-MS-MS analysis using fast atom bombardment (FAB) and packed capillary LC columns
with application of the on-column focusing technique. These methods provided the sensitivity required
to identify unknown metabolites that were present in the mature plants at only 20-230 ppb. Structural
elucidation was facilitated by use of fonofos labeled with both carbon-14 and carbon-13 in the phenyl
ring. In all three plants fonofos was converted to a glucose conjugate of thiophenoxylactic acid.
Oxidation of the glucose conjugate gave isomeric sulfoxides in all species examined. Thiophenoxylactic
acid was found esterified to malonic acid in lettuce. In beets, S-phenylcysteine was found as its
malonic acid amide. A second metabolite unique to beets was N-(malonyl)-[2[(ethoxyethylphosphi-
nothionyl)oxy]phenyl]cysteine. This novel structure was confirmed by synthesis.
KEYWORDS: Fonofos; metabolism; lettuce; wheat; beet; mass spectrometry; structure elucidation; stable
isotopes; synthesis; cysteine conjugate â-lyase
INTRODUCTION
Plant metabolism of fonofos has been studied only in Solanum
tuberosum (4). In mature tubers grown in soil treated with (ethyl-
1-14C)fonofos, 32% of the total residue was extracted as EOP.
An additional 7% of EOP resulted from treatment of bound
residues with 1 N HCl. No ETP was detected. Tubers grown
with soil containing (phenyl-U-14C)fonofos had only 3.8% of
the total residue present as methylphenyl sulfone. The majority
of the residue (86.8%) was present as water-soluble unknowns
of which 57% could be cleaved by â-glucosidase to an unknown
aglycon that could be methylated with diazomethane but that
did not cochromatograph with any of the three isomeric
hydroxyphenyl methyl sulfone derivatives.
Fonofos (O-ethyl S-phenyl ethylphosphonodithioate) is a
dithiophosphonate insecticide used to control lepidopterous
insects in corn, potatoes, and peanuts. Considerable work has
been done on the metabolism of fonofos in animals. Rat liver
microsomes convert fonofos to fonofos oxon, which then yields
O-ethyl ethylphosphonate (EOP) by hydrolysis (1). Rats treated
with (ethyl-1-14C)fonofos excreted 54.5 and 32.7% of the dose
in urine as O-ethyl ethylphosphonothioate (ETP) and EOP,
respectively (2). Rats metabolized (phenyl-U-14C)fonofos to
methylphenyl sulfone (9.9% of dose in urine) and to water-
soluble conjugates, which upon acid hydrolysis gave 3-(hy-
droxyphenyl) methyl sulfone (24.9% of dose) and 4-(hydroxy-
phenyl) methyl sulfone (23.4% of dose). The same metabolites
were found when thiophenol was administered to rats (3).
Thiophenol is thus an intermediate in the metabolism of fonofos
that is formed by hydrolysis of either fonofos or its oxon.
The use of mass spectrometry in the analysis of conjugates
and other metabolites of xenobiotics has been recently reviewed
(5). The development of continuous-flow FAB was a major
advance in the identification of polar and thermally labile
metabolites by LC-MS (6). The solvent flow requirements of
this interface matched those of packed capillary HPLC columns
with internal diameters of 0.1-0.3 mm. An injection technique
called on-column focusing has been developed that allows use
of large (10-100 µL) injection volumes with packed capillary
HPLC columns without significant deterioration of chromato-
graphic resolution (7, 8). The application of this injection
technique to the identification of metabolites of xenobiotics by
packed capillary LC-MS has been described in detail (9, 10).
* Address correspondence to this author at Syngenta, Jealotts Hill
Research Centre, Bracknell, Berkshire, RG42 6ET England (telephone 44-
1344-414-251; fax 44-1344-414-558; e-mail onibid@earthlink.com).
† Syngenta, Richmond, CA.
‡ Chiron Corp.
§ Technology Sciences Group, Inc.
# Elan Pharmaceuticals.
RF/Spectrum.
| Exelixis, Inc.
X Solano Community College.
10.1021/jf011052q CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/03/2002

Identification of Fonofos Metabolites
J. Agric. Food Chem., Vol. 50, No. 7, 2002 1923
of DL-cysteine, N-(tert-butoxycarbonyl)-2-hydroxyphenyl-, methyl ester
as an oil: 1H NMR (CDCl3, 300 MHz) δ 1.42 (s, 9H), 3.10 (dd, 1H),
3.22 (dd, 1H) or 3.16 (d of AB quartet, 2H), 3.62 (s, 3H), 4.53 (br,
1H), 5.40 (br, 1H), 6.85 (t, 1H), 6.98 (d, 1H), 7.16 (t, 1H), 7.27 (d,
1H).
Scheme 1. Synthesis of DL-Cysteine, N-(Methoxycarbonylacetyl)-
[2-[(ethoxyethylphosphinothionyl)oxy]phenyl]-, Methyl Ester
The phenol (8.56 g, 26.2 mmol), O-ethyl-ethylphosphonochlorido-
thioate (9.93 g, 57.5 mmol), Et3N (10 mL, 7.26 g, 71.9 mmol), and
(dimethylamino)pyridine (catalyst) were heated to reflux in THF (250
mL) for 4 h. The solution was cooled, diluted with water, and extracted
with EtOAc (three times). The combined extracts were washed with
brine (saturated), dried (MgSO4), filtered, and concentrated to give 14
g of an oil. Column chromatography (SiO2: hexane/EtOAc, gradient
elution) afforded 9.15 g (75%) of DL-cysteine, N-(tert-butoxycarbonly)-
[2-[(ethoxyethylphosphinothionyl)oxy]phenyl]-, methyl ester as an
oil: 1H NMR (CDCl3, 300 MHz) δ 1.31 (m, 6H), 1.41(s, 9H), 2.27 (br
sextet, 2H), 3.36 (br d, 2H), 3.55 (s, 1.5H), 3.57 (s, 1.5H), 4.18 (m,
1H), 4.38 (m, 1H), 4.57 (broad, 1H), 5.46 (br m, 1H), 7.14 (t, 1H),
7.23 (m, 2H), 7.44 (d, 1H).
HCl gas was passed through a solution of the Boc ester (5.70 g,
12.3 mmol) in ether (200 mL). The solid was collected, washed with
ether, and dried to give 2.02 g (45%) of DL-cysteine, [2-[(ethoxyeth-
ylphosphinothionyl)oxy]phenyl]-, methyl ester amine hydrochloride:
1H NMR (DMSO-d6, 300 MHz) δ 1.18 (t, 1.5H), 1.26 (m, 4.5H), 2.29
(sextet, 2H), 3.53 (s, 3H), 3.56 (m, 1H), 3.42 (m, 1H), 4.21 (m, 3H),
7.26 (m, 3H), 7.38 (d, 1H), 8.99 (br s, 3H).
Methyl malonyl chloride (0.75 g, 5.5 mmol) in CH2Cl2 (5 mL) was
added dropwise to a ice bath cooled solution of the amine salt (1.00 g,
2.73 mmol) and Et3N (1.5 mL, 1.09 g, 10.7 mmol) in CH2Cl2 (25 mL).
The solution was allowed to warm to room temperature and stirred
overnight. The mixture was diluted with NaHCO3 (saturated) and
extracted with CH2Cl2 (three times). The combined extracts were
washed with brine (saturated), dried (MgSO4), filtered, and concentrated.
Column chromatography (SiO2: hexane/EtOAc gradient elution) af-
forded 0.82 g (65%) of DL-cysteine, N-(methoxycarbonylacetyl)-[2-
[(ethoxyethylphosphinothionyl)oxy]phenyl]-, methyl ester as an oil: 1H
NMR (CDCl3, 300 MHz) δ 1.29 (m, 6H), 2.25 (sextet, 2H), 3.16 (d,
1H), 3.20 (d, 1H), 3.44 (m, 2H), 3.61 (s, 3H), 3.48 (s, 3H), 4.23 (m,
2H), 4.83 (m, 1H), 7.17 (m, 1H), 7.22 (m, 2H), 7.46 (br d, 1H), 7.78
(br t, 1H).
Soil Treatment and Planting. A sandy loam soil obtained from
Visalia, CA (57% sand, 34% silt, 9% clay, 0.8% organic matter, pH
7.4, cation exchange capacity 7.6 mequiv/100 g), was put in containers
and treated with radiolabeled fonofos dissolved in 50:50 acetone/water
at an application rate of 49 µg/cm2 of soil surface (4.7 lb/acre). The
soil was maintained at ambient temperature for 32 days in a greenhouse
in Richmond, CA, before the crops were planted. Crops were harvested
at maturity (3 months for Latuca satiVa, 3.5 months for Triticum
aestiVum, and 4 months for Beta Vulgaris).
Extraction. The mature lettuce leaves, beet leaves, beet roots, and
wheat straw were pulverized with dry ice and then extracted with 50:
50 (v/v) acetonitrile/water. The acetonitrile was removed by rotary
evaporation, and the residual aqueous solution was extracted with ethyl
acetate to remove nonpolar neutrals. The aqueous extracts were analyzed
by HPLC-RAM and were the source of all of the metabolites identified
in this study.
HPLC-RAM Analysis. HPLC-RAM analysis was done on a
Hewlett-Packard model 1090 HPLC with a Raytest Ramona model 5-LS
radioactivity monitor equipped with a solid scintillation cell. A PhaseSep
Spherisorb S5 ODS2 column (4.6 mm i.d., 250 mm length) was used
at 1 mL/min with the following gradient: 5-25% A in 60 min; 25-
55% A in 20 min; 55% A for 10 min; 55-100% A in 10 min with A
) acetonitrile containing 0.1% trifluoroacetic acid and B ) 0.1%
aqueous trifluoroacetic acid.
Metabolite Isolation. Metabolite A15 was isolated from beet leaf.
The aqueous extract was acidified with 1 M HCl and extracted with
ethyl acetate. The ethyl acetate extract was chromatographed on
preparative HPLC. A PhaseSep Spherisorb S5 ODS2 column (4.6 mm
i.d., 250 mm length) was used at 1 mL/min with the following
gradient: 15-20% A in 15 min; 20-25% A in 10 min; 25-30% A in
5 min; 30-100% A in 20 min with A ) acetonitrile containing 0.1%
trifluoroacetic acid and B ) 0.1% aqueous trifluoroacetic acid. The
It is the purpose of this study to use LC-MS and LC-MS-
MS with packed capillary HPLC and continuous-flow FAB to
identify the water-soluble metabolites of fonofos found in mature
plants. These experiments were done as part of a larger crop
rotation study done for reregistration of fonofos with the EPA.
MATERIALS AND METHODS
Chemicals. Fonofos. O-Ethyl S-phenyl-U-14C-13C6-ethylphosphon-
odithioate was synthesized by Zeneca Ag Products (Richmond, CA).
The material was prepared with a 12C/13C ratio of about 2:1 and a
specific activity of 28 Ci/mol. The synthesis has previously been
described (11). Structural confirmation was done by H NMR and EI-
GC-MS. The material was found to have a purity of 99% by GC-MS.
DL-Cysteine, N-(Ethoxycarbonylacetyl)-[2[(ethoxyethylphosphinothio-
nyl)oxy]phenyl]-, Methyl Ester. See Scheme 1.)
N-tert-Butoxycarbonylserine (20.5 g, 0.10 mol), Cs2CO3 (16.5 g, 50.6
mmol), and CH3I (12 mL, 27.3 g, 0.192 mol, 1.9 equiv) were stirred
in DMF (250 mL) overnight. The mixture was then diluted with water,
and the product was extracted with EtOAc (three times). The combined
extracts were washed with brine (saturated), dried (MgSO4), filtered,
and concentrated to give 17.05 g (78% isolated) of N-tert-butoxycar-
bonylserine methyl ester as an oil: 1H NMR (CDCl3, 300 MHz) δ
1.45 (s, 9H), 2.79 (br t, 1H), 3.78 (s, 3H), 3.91 (m, 2H), 4.4 (br d, 1H),
5.75 (br d, 1H).
Methane sulfonyl chloride (6.0 mL, 77.5 mmol) in CH2Cl2 (25 mL)
was added dropwise to an ice-cold solution of N-tert-butoxycarbon-
ylserine methyl ester (15.3 g, 69.8 mmol) and Et3N (10.7 mL, 7.70 g,
77.0 mmol) in CH2Cl2 (150 mL). The ice bath was removed, and the
solution was stirred for 2.5 h, then partitioned with brine (saturated)
and extracted with CH2Cl2 (three times), dried (MgSO4), filtered, and
concentrated to give 19.56 g (94% isolated) of the mesylate of N-tert-
butoxycarbonylserine methyl ester: 1H NMR (CDCl3, 300 MHz) δ 1.46
(s, 9H), 3.03 (s, 3H), 3.81 (s, 3H), 4.48 (m, 1H), 4.59 (m, 2H), 5.47
(br d, 1H).
Et3N (8.0 mL, 5.80 g, 57.5 mmol) in THF (25 mL) was added
dropwise to a solution of the mesylate (15.23 g, 51.27 mmol) in THF
(250 mL) and allowed to stir for 6 h, diluted with brine (saturated),
and extracted with CH2Cl2 (two times). The combined extracts were
dried (MgSO4), filtered, and concentrated to give 11.4 g of crude N-tert-
butoxycarbonyldehydroalanine methyl ester. This crude ester was
combined with 2-hydroxythiophenol (6.53 g, 51.8 mmol), and Et3N
(1.40 mL, 1.01 g, 10.0 mmol) and stirred in degassed CH3OH (200
mL), under N2, for 3 days. The solution was diluted with brine
(saturated) and extracted with EtOAc (three times). The combined
extracts were dried (MgSO4), filtered, and concentrated. Column
chromatography (SiO2: hexane/EtOAc, gradient elution) gave 8.56 g

1924 J. Agric. Food Chem., Vol. 50, No. 7, 2002
Onisko et al.
fractions containing metabolite A15 were methylated with excess
ethereal diazomethane. The methylated sample was further purified
before LC-MS analysis by HPLC using 45:55 (v/v) acetonitrile/0.1%
aqueous trifluoroacetic acid.
Metabolites A4, A5, and A14 were isolated from lettuce leaf. A
Phenomenex Spherisorb 5 ODS2 column (10 mm i.d., 250 mm length)
was used at 3 mL/min to purify the aqueous plant extract using the
following gradient: 0-50% A in 25 min; 50% A for 10 min; 50-
100% A in 5 min with A ) acetonitrile containing 0.1% trifluoroacetic
acid and B ) 0.1% aqueous trifluoroacetic acid. This gave an A4/A5
fraction and an A14 fraction. The A4/A5 fraction was purified by HPLC
using 7.5:92.5 (v/v) acetonitrile/water to give fractions A4 and A5.
Metabolite A4 was methylated with excess ethereal diazomethane and
analyzed by LC-MS. Metabolite A5 was methylated with excess
ethereal diazomethane and was further purified before LC-MS analysis
by HPLC using 12:88 (v/v) acetonitrile/0.1% aqueous trifluoroacetic
acid. The A14 fraction from the preparative chromatography was further
purified by HPLC (20:70 acetonitrile/water) and then methylated with
excess ethereal diazomethane. The methylated material was further
purified before LC-MS analysis by HPLC using 40:60 (v/v) acetonitrile/
water.
Figure 1. Molecular ion region from the electron impact mass spectrum
of [14C-UL-phenyl]-[13C -phenyl]-O-ethyl-S-phenylethyl-phosphonodithioate.
6
RESULTS AND DISCUSSION
To identify metabolites of fonofos in mature plants, U-14C-
13C6-labeled fonofos (O-ethyl S-phenyl-U-14C-13C6-ethylphospho-
nodithioate) was synthesized. The label incorporated 14C to allow
quantitation and to monitor purification of metabolites. The
purpose of the 13C6 -label was (1) to provide a recognizable
isotope pattern to readily distinguish fonofos metabolites from
unlabeled natural products and (2) to allow tandem mass
spectrometric experiments of both labeled and unlabeled precur-
sor ions to aid in structure elucidation. Figure 1 shows the
molecular ion region of the mass spectrum obtained by electron
impact GC-MS of the U-14C-13C6-labeled fonofos used in the
metabolism studies. The most abundant peak was observed at
m/z 246 corresponding to fonofos, where all carbons are carbon-
12 (12C10H15POS2). The next most abundant peak was observed
at m/z 252 (13C612C4H15POS2). This peak is from the 13C6-
phenyl-labeled material. Despite the high specific activity of
this material (28 Ci/mol), ions corresponding to molecules of
fonofos containing four, five, and six atoms of carbon-14 (at
m/z 254, 256, and 258) are of low abundance (4-8% of the
intensity of the ion of m/z 246) and were not useful for the
recognition of fonofos-related ions of unknown metabolites.
Soil treated with [U-14C-13C6]fonofos was kept at ambient
temperature for 32 days before seeds of lettuce (Latuca satiVa),
beet (Beta Vulgaris), or wheat (Triticum aestiVum) were planted.
The plants were grown to maturity and the lettuce leaves, beet
leaves, beet roots, and wheat straw were extracted with
acetonitrile/water. HPLC-RAM analysis of the various extracts
gave the radiochromatograms shown in Figure 2. Six major
and a variety of minor radioactive peaks (named A1-A15) were
found. All of the metabolites were more polar than fonofos,
which was not detected in mature plant tissue.
Metabolite A12 was isolated from beet root. The aqueous extract
was acidified with 1 M HCl and extracted with ethyl acetate. The ethyl
acetate extract was chromatographed on a PhaseSep Spherisorb S5
ODS2 column (4.6 mm i.d., 250 mm length) at 1 mL/min with the
following gradient: 0% A for 5 min; 0-50% A in 20 min; 50% A for
5 min; 50-100% A in 5 min with A ) acetonitrile containing 0.1%
trifluoroacetic acid and B ) 0.1% aqueous trifluoroacetic acid. The
fractions containing metabolite A12 were methylated with excess
ethereal diazomethane before LC-MS analysis.
Metabolite A10 was isolated from wheat straw. The aqueous extract
was chromatographed on preparative HPLC using the same column
and method used to chromatograph the lettuce extract. The A10 fraction
was further purified by HPLC using 25:75 (v/v) acetonitrile/0.1%
aqueous trifluoroacetic acid and then again using 15:85 (v/v) acetoni-
trile/0.1% aqueous trifluoroacetic acid. Metabolite A10 was methylated
with excess ethereal diazomethane and was further purified before LC-
MS analysis by HPLC using 20:80 (v/v) acetonitrile/water.
NMR and Mass Spectrometry. NMR spectra were obtained on a
General Electric model QE-300 Fourier transform spectrometer.
Packed capillary LC flow FAB was done on a Finnigan-MAT model
TSQ 700 triple-quadrupole mass spectrometer (San Jose, CA) equipped
with a flow FAB source. Mobile phases were prepared that contained
acetonitrile, aqueous 0.1% trifluoroacetic acid, and 5% glycerol. The
mobile phase was pumped with a Perkin-Elmer model 250 HPLC
(Norwalk, CT) to a splitter connected to (1) an HPLC column (1 mm
diameter by 200 mm long ABI Spheri-5 RP-18, 5 µm) that flowed to
waste and (2) a Rheodyne model 8125 injector (Cotati, CA) with a 5.0
µL sample loop. Prior to injection, purified methylated metabolite
samples (10-200 ng/5 µL injection) were concentrated to dryness and
then redissolved in LC-MS mobile phase that had been diluted 2-fold
with water to achieve solvent focusing. A 0.322 mm i.d. × 150 mm
capillary HPLC column packed with Spherisorb ODS-2 (3 µm) from
LC Packings (San Francisco, CA) was connected to the outlet of the
Rheodyne injector. A flow rate of 8-10 µL/min through the packed
capillary HPLC column was achieved by adjusting the total flow from
the Perkin-Elmer pump from 0.15 to 0.2 mL/min. The packed capillary
HPLC column effluent flowed through an ABI model 785A UV detector
(Foster City, CA) that was operated at 210 nm and was equipped with
a 60 nL flow cell, model UZ-View, from LC Packings. The flow from
the UV detector was connected to the flow FAB source of the TSQ
700. The source was operated at 20-30 °C and bombarded with 6-7
kV xenon atoms. Source stability was achieved with the use of a Bausch
and Lomb model Stereo 4 binocular microscope. MS-MS experiments
were done with 0.5 mTorr of argon as collision gas. Ions were
accelerated prior to collision-induced dissociation (CID) by applying
a 20 V potential difference between the first quadrupole and the collision
cell. Scans of 1 s duration were acquired in all cases.
Metabolite A8 (see Figure 2) cochromatographed with
methylphenyl sulfone. The other labeled metabolites were
identified by LC-MS. Extensive purification was required before
LC-MS analysis due to the low concentration of metabolites in
plants grown in soil containing biologically relevant amounts
of fonofos. For example, the maximum application rate used in
agriculture (4 lb/acre) corresponds to a concentration of fonofos
in the top 7 cm of soil of ∼5 ppm. At maturity the plants
contained only 0.3-1.7% of the applied radioactivity. Concen-
trations in the plant tissues used to isolate metabolites A4, A15,
A5, A14, and A12 were only 20, 40, 60, 70, and 230 ppb,
respectively.
FAB gave significantly larger MH+ intensities from the
methylated samples than for the free acids. Consequently, all
of the metabolites were analyzed as the free acid and as their

Identification of Fonofos Metabolites
J. Agric. Food Chem., Vol. 50, No. 7, 2002 1925
Figure 3. CID spectra of m/z 470 (top) and 464 (bottom) from the LC-
MS-MS analysis of methylated metabolite A15.
in both CID spectra and therefore contain none of the phenyl
carbons.
To determine the number of sulfur atoms in the CID fragment
ions, methylated metabolite A15 was analyzed by LC-MS-MS
observing product ions of the 34S isotope peak at m/z 466.
Fragment ions containing no sulfurs should appear at the same
m/z value in CID spectra of either m/z 464 or 466. Fragments
containing one sulfur will produce two equal intensity peaks
separated by 2 amu in the CID spectrum of m/z 466, because
the precursor ions contain the 34S atom in either one of two
possible positions. Fragment ions that contain both sulfurs will
produce a single peak in the CID spectrum of m/z 466, but
because the fragment contains the 34S atom of the precursor, it
will have an m/z ratio 2 amu higher than the analogous ion in
the CID spectrum of the m/z 464 ion. The presence of 18O and
13C2 complicates the analysis somewhat. In the CID spectrum
of m/z 466, the observed ratio of intensities of m/z x + 2 to m/z
x is shown in Figure 4 for nine fragment ions (x ) 109, 125,
137, 170, 202, 211, 233, 261, and 364). Also shown in Figure
4 are the theoretical intensity ratios for the nine ions of interest
assuming zero, one, or two sulfurs per ion. By comparing the
observed-to-predicted ratios it is clear that the fragment ions of
m/z 170 and 202 contain no sulfur atoms, the fragment ions of
m/z 109, 125, 137, and 211 contain one sulfur atom, and the
fragment ions of m/z 233, 261, and 364 contain two sulfur atoms.
The CID fragment ion from A15 of m/z 125 contained one
sulfur and all six phenyl carbons, suggesting the empirical
formula of C6H5OS and that metabolite A15 contained hydroxy-
thiophenol as a substructure. The fragment ions of m/z 109 and
137 contained one sulfur atom and no phenyl carbons. Further-
more, ions of m/z 109 and 137 are found as fragment ions of
fonofos (electron impact data, not shown), which suggests the
substructures C2H5PSOH and C2H5PSOC2H5 for the ions of m/z
109 and 137. Assembling these substructures provided the
Figure 2. Radiochromatogram (14C, y-axis) of lettuce leaf (bottom), beet
root, beet leaf, and wheat straw (top) extracts by gradient HPLC. The
x-axis is time after injection from 0 to 100 min.
methyl esters. LC-MS analysis of native metabolite A15 gave
an MH+ ion of m/z 436. Metabolite A15 was then methylated
with diazomethane and examined by packed capillary LC-MS.
The corresponding mass spectrum showed an MH+ ion of m/z
464, which indicates that the peak is a dimethyl ester of the
native metabolite. The MH+ ion of the methylated metabolite
showed the same isotopic pattern as for the starting fonofos,
demonstrating the presence of the intact phenyl group in the
unknown metabolite. Furthermore, the intensity of the MH+ +
2 ion relative to MH+ (11.8 ( 0.6% for the mean ( SD of
three independent observations) suggests the presence of two
sulfur atoms in metabolite A15.
Subsequent analysis of methylated metabolite A15 by LC-
MS-MS observing product ions of m/z 464 resulted in the CID
spectrum shown in Figure 3 (bottom). The fragment ion of m/z
109 was first attributed to C6H5S; however, no reasonable
structure for A15 could be assembled that contained this
substructure. Next the methylated metabolite was analyzed by
LC-MS-MS observing product ions of the 13C6 isotope peak at
m/z 470. This gave the CID spectrum shown in the top of Figure
3. By comparing the two CID spectra in Figure 3 it is possible
to determine which fragment ions contain the phenyl ring. For
example, the CID fragments at m/z 125, 211, 233, 261, 263,
364, and 432 contain all six carbons of the phenyl group because
these ions are observed at m/z values 6 amu higher in the
analogous CID spectrum of the 13C6-labeled MH+ ion. The ions
of m/z 109, 137, 170, and 202 were found at the same m/z value

1926 J. Agric. Food Chem., Vol. 50, No. 7, 2002
Onisko et al.
Figure 6. CID spectrum of m/z 313 from the LC-MS-MS analysis of
methylated metabolite A14.
Scheme 2. Proposed Mechanism for Biosynthesis of Metabolite A15
Figure 4. Predicted (solid lines) vs observed (b) ion intensity ratios for
ions of m/z 109 (C H OPS ), 125 (C H OS ), 137 (C H OPS ), 170 (C H -
2
6
x
6
5
x
4
10
x
7 8
NO S ), 202 (C H NO S ), 211 (C H NO S ), 233 (C H O PS ), 261
4
x
8
12
5
x
10 13
2
x
8
10
2
x
(C H O PS ), and 364 (C H NO PS ) from the LC-MS-MS product
10 14
2
x
14 23
4
x
ion analysis of m/z 466 of methylated metabolite A15 where x ) 0
(bottom), 1 (middle), or 2 (top). Predicted ion intensity ratio ) [(a + b +
c)/((a(no. of C in ion/18) + b(no. of O in ion/7) + c(x/2))] − 1, where a )
0.005(1.08 × 18)2, b ) 0.02(7), and c ) 4.21(2).
Figure 5. CID fragmentation scheme for methylated metabolite A15.
structure proposed for the dimethyl ester of metabolite A15.
Figure 5 shows this structure as well as the origin of the major
CID fragmentation ions. Further work was needed to determine
the hydroxythiophenol substitution pattern in metabolite A15.
Due to biosynthetic mechanistic reasons the 1,2-substituted
isomer was considered to be most likely. A proposed pathway
is shown in Scheme 2. Metabolite A15 could arise via
glutathione addition to a benzene epoxide, rearrangement, and
elimination of H2S to give the glutathione conjugate of the
fonofos oxon isomer. Subsequent hydrolysis to lose glycine and
glutamate followed by malonylation would give metabolite A15.
The proposed rearrangement occurs via a five-member transition
state and results in biosynthesis of a 1,2-substituted hydroxy-
thiophenol. Consequently, the 1,2-substituted isomer of A15 was
synthesized (see Scheme 1). The synthetic material cochro-
matographed with the methylated metabolite on HPLC. Fur-
thermore, the CID spectrum obtained by LC-MS-MS analysis
of synthetic A15M was nearly indistinguishable from the
analogous CID spectra of the metabolite, demonstrating that
the proposed structure is correct.
contained fragment ions of m/z 109, 135, and 195. This
suggested the structure of a malonate ester of thiophenoxylactic
acid.
Packed capillary LC-MS analysis of methylated metabolite
A12 showed an MH+ ion of m/z 312. Analysis of methylated
peak A12 by LC-MS-MS was done with observation of product
ions of m/z 312. The CID spectrum obtained (see Figure 7)
contained fragment ions of m/z 101, 118, 135, 195, 202 (as in
A15), 252, and 280, suggesting the malonic acid amide of
S-phenylcysteine. This structure and the origin of the major CID
fragment ions are also shown in Figure 7.
Packed capillary LC-MS analysis of methylated metabolite
A14 showed an MH+ ion of m/z 313. Analysis of methylated
peak A14 by LC-MS-MS was done with observation of product
ions of m/z 313. The CID spectrum (see Figure 6) obtained
Packed capillary LC-MS analysis of methylated metabolite
A10 showed an MH+ ion of m/z 375. Analysis of methylated

Identification of Fonofos Metabolites
J. Agric. Food Chem., Vol. 50, No. 7, 2002 1927
Figure 9. CID spectrum of m/z 391 from the LC-MS-MS analysis of
Figure 7. CID spectrum of m/z 312 from the LC-MS-MS analysis of
methylated metabolite A4.
methylated metabolite A12.
Scheme 3. Proposed Pathway for Metabolism of Fonofos
Figure 8. CID spectrum of m/z 375 from the LC-MS-MS analysis of
methylated metabolite A10.
peak A10 by LC-MS-MS was done with observation of product
ions of m/z 375. The CID spectrum obtained contained fragment
ions of m/z 109, 135, 153, 195, and 213, suggesting the
glucoside of thiophenoxylactic acid. This structure and the origin
of the major CID fragment ions are shown in Figure 8.
Packed capillary LC-MS analysis of methylated metabolites
A4 and A5 showed both to have an MH+ ion of m/z 391.
Analysis of methylated peak A4 and methylated peak A5 by
LC-MS-MS was done with observation of product ions of m/z
391. The CID spectra obtained were nearly indistinguishable
and contained fragment ions of m/z 103, 127, and 229,
suggesting isomers of molecular weight 16 amu higher than
that of methylated metabolite A10. Oxidation of metabolite A10
with hypochlorite gave material that comigrated on HPLC with
metabolites A4 and A5. This suggested that metabolites A4 and
A5 are sulfoxides of metabolite A10. These structures and the
origin of the major CID fragment ions are shown in Figure 9.
Furthermore, methylation gave only monomethyl derivatives for
metabolites A4 and A5. (The alternative isomers with the phenyl
ring oxidized to a phenol should have methylated to give
dimethyl derivatives of molecular weight 404.) Finally, the
analysis of methylated peak A4 and methylated peak A5 by
LC-MS-MS was done with observation of product ions of m/z
229, which corresponds to the protonated aglycon. The CID
spectrum obtained contained fragment ions of m/z 71, 103, 109,
125, 127, 169, and 197. The fragment of m/z 109 (C6H5S) is
much more likely to arise from the proposed sulfoxides than
from the alternative phenolic isomers.
phenyl sulfide, methyl phenyl sulfoxide, and ultimately methyl-
phenyl sulfone.
The cysteine and lactic acid derivatives found in this study
(A4, A5, A10, A12, and A14) appear to be typical plant
metabolites of a glutathione conjugate (12). The metabolic
pathway proposed for formation of the fonofos metabolites
identified in this paper is shown in Scheme 3 and shows how
these metabolites could be formed from a glutathione conjugate
of thiophenol. Scheme 3 also shows an alternative route to these
metabolites via hydrolysis of fonofos to thiophenol and then
synthesis of the cysteine conjugate of thiophenol. The interme-
diacy of thiophenol is supported by the identification of
Prior to this study, the metabolism of fonofos in plants has
been studied only in Solanum tuberosum (4). Fonofos is reported
to be metabolized to thiophenol, which is converted to methyl

1928 J. Agric. Food Chem., Vol. 50, No. 7, 2002
Onisko et al.
(2) McBain, J. B.; Hoffman, L. J.; Menn, J. J.; Casida, J. E. Dyfonate
Metabolism Studies II. Metabolic Pathway of O-Ethyl S-Phenyl
Ethylphosphonodithioate in Rats. Pestic. Biochem. Physiol. 1971,
1, 356-365.
(3) McBain, J. B.; Menn, J. J. S-Methylation, oxidation, hydroxy-
lation and conjugation of thiophenol in the rat. Biochem.
Pharmacol. 1969, 18, 2282-2285.
(4) McBain, J. B.; Hoffman, L. J.; Menn, J. J. Metabolic Degradation
of Dyfonate in Potato Plants. J. Agric. Food Chem. 1970, 18,
1139-1144.
(5) Burlingame, A. L.; Boyd, R. K.; Gaskell, S. J. Mass Spectrom-
etry. Anal. Chem. 1998, 70, 647R-716R.
(6) Caprioli, R. M.; Tomer, K. Liquid Chromatography/Mass
Spectrometry. In Continuous-Flow Fast Atom Bombardment
Mass Spectrometry; Caprioli, R. M., Ed.; Wiley: New York,
1990.
(7) Ling, B. L.; Baeyens, W.; Dewaele, C. Comparative study of
“on-column focusing” applied to micro-LC and conventional LC.
J. Microcolumn Sep. 1992, 4 (1), 17-22.
(8) Brunmark, P.; Dalene, M.; Sango, C.; Skarping, G.; Erlandsson,
P.; Dewaele, C. Determination of toluene diisocyanate in air
using micro liquid chromatography with UV-detection. J. Mi-
crocolumn Sep. 1991, 3, 371-375.
(9) Onisko, B. C.; Barnes, J. P.; Staub, R. E.; Walker, F. H.;
Kerlinger, N. Metabolism of Cycloate in Radish Leaf: Metabolite
Identification by Packed Capillary Flow Fast Atom Bombard-
ment Tandem Mass Spectrometry. Biol. Mass Spectrom. 1994,
23, 626-636.
methylphenyl sulfone, formed most likely by methylation and
oxidation of thiophenol. Synthesis of the cysteine conjugate of
thiophenol could be accomplished by addition of pyruvate and
NH3 to thiophenol by a cysteine conjugate â-lyase (EC 4.4.1.13)
operating in reverse (C-S synthesis, not C-S cleavage).
Cysteine conjugate â-lyases have been well documented in the
rat (13) and in plants (14). To our knowledge, this enzyme has
never been shown to act in the biosynthesis of cysteine
conjugates. Study of the metabolism of [34S2]fonofos could
prove whether the glutathione- or the thiophenol-related pathway
proposed in Scheme 3 is operative in plants and provide
evidence for a C-S lyase acting in the biosynthesis of cysteine
conjugates.
ABBREVIATIONS USED
LC-MS, liquid chromatography-mass spectrometry; GC-MS,
gas chromatography-mass spectrometry; FAB, fast atom
bombardment; EOP, O-ethyl ethylphosphonate; ETP, O-ethyl
ethylphosphonothioate; HPLC, high-pressure liquid chroma-
tography; DMF, dimethylformamide; NMR, nuclear magnetic
resonance; THF, tetrahydrofuran; RAM, radioactivity monitor;
UV, ultraviolet; TIC, total ion current; CID, collision-induced
dissociation; GSH, glutathione.
SAFETY
(10) Onisko, B. C. Advances in Metabolite Identification. In Eighth
International Congress of Pesticide Chemistry Options 2000;
Ragsdale, N. N., Kearney, P. C., Plimmer, J. R., Eds.; American
Chemical Society: Washington, DC, 1995.
(11) Kalbfeld, J.; Pitt, H. M.; Hermann, D. A. Synthesis of O-Ethyl
S-Phenyl-14C(U)-Ethylphosphonodithioate. J. Labelled Com-
pounds 1969, 5, 351-354.
(12) Lamoureux, G. L.; Rusness, D. G. The Role of Glutathione and
Glutathione-S-Transferases in Pesticide Metabolism, Selectivity,
and Mode of Action in Plants and Insects. In Coenzymes and
Cofactors; Dolphin, D., Poulson, R., Avramovic, O., Eds.;
Wiley: New York, 1989; pp 154-192.
(13) Perry, S. J.; Schofield, M. A.; MacFarlane, M.; Lock, E. A.;
King, L. J.; Gibson, G. G.; Goldfarb, P. S. Isolation and
expression of a cDNA coding for rat kidney cytosolic cysteine
conjugate beta-lyase. Mol. Pharmacol. 1993, 43, 660-665.
(14) Nock, L. P.; Mazelis, M. The C-S lyases of higher plants:
preparation and properties of homogeneous alliin lyase from
garlic (Allium satiVum). Arch. Biochem. Biophys. 1986, 249, 27-
33.
Diazomethane is an explosive (use safety shield), insidious
poison (a well-ventilated hood is absolutely necessary) and a
strong irritant, which does not cause discernible reaction at the
time of contact but later, even in minute amounts, produces an
inflammatory response.
N-Methyl-N′-nitro-N-nitrosoguanidine is the starting material
for the preparation of diazomethane. It is a highly flammable,
toxic mutagen that may cause cancer, may cause heritable
genetic damage, is a possible teratogen, is toxic if swallowed,
is harmful by inhalation, and is irritating to the eyes, respiratory
system, and skin.
ACKNOWLEDGMENT
We gratefully acknowledge Jules Kalbfeld and Angel Tolentino
for preparation of the radiolabeled fonofos used in this study.
We also thank Lydia Chang for NMR interpretation.
LITERATURE CITED
Received for review August 7, 2001. Revised manuscript received
December 12, 2001. Accepted December 12, 2001.
(1) McBain, J. B.; Yamamoto, I.; Casida, J. E. Thionophosphorus
insecticides: Mechanism of activation and deactivation of
Dyfonate by rat liver microsomes. Life Sci. 1971, 10, 947-954.
JF011052Q