Oxidative bioactivation of methamidophos insecticide: Synthesis of N- hydroxymethamidophos (a candidate metabolite) and its proposed alternative reactions involving N → O rearrangement or fragmentation through a metaphosphate analogue

Article abstract of DOI:10.1021/tx9701135

The systemic insecticide methamidophos, MeO(MeS)P(O)NH₂, is a very weak inhibitor of acetylcholinesterase (ACHE) in vitro relative to in vivo suggesting bioactivation. This hypothesis is supported by finding that brain AChE inhibition and poisoning signs from methamidophos are greatly delayed in mice and houseflies pretreated with oxidase inhibitors in an order for effectiveness of methimazole > N-benzylimidazole >> piperonyl butoxide. In contrast, the order for delaying parathion-induced AChE inhibition and toxicity is N-benzylimidazole >> piperonyl butoxide or methimazole, suggesting that different oxidases are involved in methamidophos and parathion activation. N-Hydroxylation is examined here as an alternative to the controversial S-oxidation proposed earlier for methamidophos activation. N-Hydroxymethamidophos [MeO(MeS)P(O)NHOH], synthesized by coupling MeO(MeS)P(O)Cl and Me₃-SiNHOSiMe₃ followed by desilylation, is unstable at pH 7.4 (t( 1/4 ) = 10 min at 37 °C) with decomposition by two distinct and novel mechanisms. The first mechanism (A) is N → O rearrangement to MeO(MeS)P(O)ONH₂ and then hydrolysis to MeO(MeS)P(O)OH, a sequence also established in the analogous series of (MeO)₂P(O)NHOH → (MeO)₂P(O)ONH₂ → (MeO)₂P-(O)OH. The second mechanism (B) is proposed to involve tautomerism to the phosphimino form [MeO(MeS)P(OH)=NOH] that eliminates MeSH forming a metaphosphate analogue [MeOP(O)=NOH] trapped by water to give MeO(HO)P(O)NHOH that undergoes the N → O rearrangement as above and hydrolysis to MeOP(O)(OH)₂. As a metaphosphate analogue, the metaphosphorimidate generated from MeO(MeS)P(O)NHOH in aqueous ethanol yields MeOP(O)(OH)₂ and MeO(EtO)P(O)OH in the same ratio as the solvents on a molar basis. Reactions of the N- and O-methyl derivatives of MeO(MeS)P(O)NHOH and (MeO)₂P(O)NHOH are consistent with proposed mechanisms A and B. N-Hydroxymethamidophos is less potent than methamidophos as an AChE inhibitor and toxicant possibly associated with its rapid hydrolysis. Bioactivation of methamidophos via a metaphosphate analogue would directly yield a phosphorylated and aged AChE resistant to reactivating agents, an intriguing hypothesis worthy of further consideration.

Full text of DOI:10.1021/tx9701135

26
Chem. Res. Toxicol. 1998, 11, 26-34
Oxid a tive Bioa ctiva tion of Meth a m id op h os In secticid e:
Syn th esis of N-Hyd r oxym eth a m id op h os (A Ca n d id a te
Meta bolite) a n d Its P r op osed Alter n a tive Rea ction s
In volvin g N f O Rea r r a n gem en t or F r a gm en ta tion
th r ou gh a Meta p h osp h a te An a logu e
Mahmoud Mahajna and J ohn E. Casida*
Environmental Chemistry and Toxicology Laboratory, Department of Environmental Science,
Policy and Management, University of California, Berkeley, California 94720-3112
Received J uly 1, 1997
The systemic insecticide methamidophos, MeO(MeS)P(O)NH2, is a very weak inhibitor of
acetylcholinesterase (AChE) in vitro relative to in vivo suggesting bioactivation. This hypothesis
is supported by finding that brain AChE inhibition and poisoning signs from methamidophos
are greatly delayed in mice and houseflies pretreated with oxidase inhibitors in an order for
effectiveness of methimazole > N-benzylimidazole . piperonyl butoxide. In contrast, the order
for delaying parathion-induced AChE inhibition and toxicity is N-benzylimidazole . piperonyl
butoxide or methimazole, suggesting that different oxidases are involved in methamidophos
and parathion activation. N-Hydroxylation is examined here as an alternative to the
controversial S-oxidation proposed earlier for methamidophos activation. N-Hydroxymetha-
midophos [MeO(MeS)P(O)NHOH], synthesized by coupling MeO(MeS)P(O)Cl and Me3-
SiNHOSiMe3 followed by desilylation, is unstable at pH 7.4 (t1/2 ) 10 min at 37 °C) with
decomposition by two distinct and novel mechanisms. The first mechanism (A) is N f O
rearrangement to MeO(MeS)P(O)ONH2 and then hydrolysis to MeO(MeS)P(O)OH, a sequence
also established in the analogous series of (MeO)2P(O)NHOH f (MeO)2P(O)ONH2 f (MeO)2P-
(O)OH. The second mechanism (B) is proposed to involve tautomerism to the phosphimino
form [MeO(MeS)P(OH)dNOH] that eliminates MeSH forming a metaphosphate analogue
[MeOP(O)dNOH] trapped by water to give MeO(HO)P(O)NHOH that undergoes the N f O
rearrangement as above and hydrolysis to MeOP(O)(OH)2. As a metaphosphate analogue,
the metaphosphorimidate generated from MeO(MeS)P(O)NHOH in aqueous ethanol yields
MeOP(O)(OH)2 and MeO(EtO)P(O)OH in the same ratio as the solvents on a molar basis.
Reactions of the N- and O-methyl derivatives of MeO(MeS)P(O)NHOH and (MeO)2P(O)NHOH
are consistent with proposed mechanisms A and B. N-Hydroxymethamidophos is less potent
than methamidophos as an AChE inhibitor and toxicant possibly associated with its rapid
hydrolysis. Bioactivation of methamidophos via a metaphosphate analogue would directly yield
a phosphorylated and aged AChE resistant to reactivating agents, an intriguing hypothesis
worthy of further consideration.
ported by in vitro studies with a microsomal activation
system and AChE (6). One goal of this research is to
reexamine the possible bioactivation of 1 as an AChE
inhibitor in mammals and insects.
In tr od u ction
Methamidophos (1) (Scheme 1) has the simplest struc-
ture of the major organophosphorus insecticides (1). It
is highly soluble in water and readily translocated as a
systemic to protect plants from insect attack. Compound
1 is a weak acetylcholinesterase (AChE)1 inhibitor in
vitro (2) relative to its action in vivo in mammals and
insects (3). This apparent anomaly is rationalized by
invoking slowly accumulating inhibition allowed by the
persistence of 1 (4) or bioactivation to a more potent
AChE inhibitor (3, 5), a proposition which is not sup-
Sch em e 1. Ca n d id a te Mech a n ism s for
Bioa ctiva tion of Meth a m id op h os (1) by S- a n d
N-Oxid a tion
* Corresponding author. Tel: 510-642-5424. Fax: 510-642-6497.
E-mail: ectl@nature.berkeley.edu.
1
Abbreviations: AChE, acetylcholinesterase; CI, chemical ioniza-
tion; FMO, flavin-containing monooxygenase; IC50, concentration for
50% inhibition; KD, knockdown; NBI, N-benzylimidazole; nosylate,
p-nitrobenzenesulfonate derivative; NsCl, nitrobenzenesulfonyl chlo-
ride; PB, piperonyl butoxide; Me, methyl; Et, ethyl; Bu, butyl; Bn,
benzyl.
S0893-228x(97)00113-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/01/1998

N-Hydroxymethamidophos Chemistry and Action
Chem. Res. Toxicol., Vol. 11, No. 1, 1998 27
cloth. Assays involved incubation of 1 mg of protein (16) with
1 mM acetylthiocholine and 1 mM 5,5′-dithiobis(2-nitrobenzoic
acid) in pH 7.4 phosphate buffer as above (5 mL final volume)
for 15 min at 37 °C. AChE activity was determined as the
absorbance at 412 nm for the 5-thio-2-nitrobenzoate liberated
on reaction of thiocholine with the disulfide compared to that
for
a control with eserine sulfate (0.5 mM) to inhibit all
nonenzymatic thiocholine release (17). For in vivo inhibition
studies, the toxicant-treated animals and the corresponding
controls (carrier solvent only) were sacrificed by cervical disloca-
tion for mice or freezing on dry ice for houseflies with assay of
AChE activity as above. For in vitro inhibition studies, enzyme
from untreated mice was incubated with the candidate inhibitor
for 30 min at 37 °C, followed by addition of acetylthiocholine
and 5,5′-dithiobis(2-nitrobenzoic acid) and incubation for 15 min,
and then addition of eserine sulfate (0.5 mM) and absorbance
measurement at 412 nm. Other enzymes used were bovine
erythrocyte AChE (Sigma, St. Louis, MO) assayed as above and
R-chymotrypsin determined with p-nitrophenyl acetate as the
substrate (18), in each case with a 30-min incubation of enzyme
and inhibitor at 37 °C before assay of residual enzyme activity.
F igu r e 1. O,S-Dimethyl and O,O-dimethyl phosphoramidates
(1 and 2), N-hydroxyphosphoramidates (3 and 7), O-amino
phosphates (11 and 12) (proposed as transient intermediates),
and their derivatives (4-6, 8-10, and 13-15) examined.
P ossible Bioa ctiva tion of 1 a s a n ACh E In h ibitor a n d
Meta bolism in in Vitr o System s. As a control, mouse brain
homogenate (1 mg of protein) was incubated with 1 (0-100 µM)
in 100 mM phosphate buffer (pH 7.4) (1 mL) for 60 min at 37
°C; the inhibitor concentrations were selected to give ap-
proximately 0%, 10%, 20%, 50%, and 70% inhibition of AChE
activity. To test for possible bioactivation, this procedure was
varied by adding fresh mouse liver microsomes (0 or 1 mg of
protein) and NADPH (0 or 1 mM final concentration) for
coincubation with the AChE source. Residual AChE activity
was then compared in the presence and absence of microsomes
and/or NADPH to determine their effect on the inhibitory
activity of 1. As an alternative candidate oxidase for activation,
mouse flavin-containing monooxygenase (FMO 1) (as a recom-
binant enzyme expressed in Escherichia coli)2 was examined
in the same way as above, replacing the 1 mg of microsomal
protein by 4 µg of FMO 1.
Possible enzymatic metabolism of 1 was examined in two
ways. First, [14CH3S]1 (supplied by Allen Rose, Valent USA,
Walnut Creek, CA) (50 mCi/mmol) (1 µg) was incubated with
NADPH (1 mg) and mouse microsomes (1 mg of protein) in 100
mM sodium phosphate buffer (pH 7.4, 1 mL) or purified FMO
1 (4 µg) in 130 mM phosphate (pH 7.6) containing 0.13 mM
EDTA. After 60 min at 37 °C, the solutions were passed through
microspin filters (0.45-µm nylon; Lida Manufacturing, Kenosha,
WI) and 100-µL aliquots of the filtrates were analyzed by adding
unlabeled 1 (20 µg) and HPLC using an RP-18 reverse-phase
column, followed by liquid scintillation counting (19). Second,
1 (1-10 mM) was compared with methimazole [1 mM, a
preferred substrate for FMO 1 (20)] added to the purified
enzyme (4 µg/mL) by monitoring oxidation of NADPH (1 mM)
as decreased absorbance at 340 nm (21).
Chemical models served as the basis for proposing
bioactivation of 1 to 1 sulfoxide (Scheme 1) as a potent
phosphorylating agent and AChE inhibitor (5, 7). Syn-
thesis of 1 sulfoxide has been claimed (5, 8) and refuted
(9-11), but this activation hypothesis remains in doubt.
Structure-activity relationships establish that both the
PSMe and PNH2 moieties are required for the high
potency of 1 in vivo and therefore probably contribute
uniquely to its purported bioactivation mechanism (3).
N-Hydroxylation is an alternative hypothesis for bioac-
tivation of 1 based on analogous reactions with N-2-
fluorenylacetamide (12), acetaminophen and analogues
(13), and ethyl carbamate (14, 15). One way to test this
hypothesis is to synthesize N-hydroxymethamidophos (3)
(Scheme 1) and evaluate its biological activity and
reactivity relative to appropriate analogues. This report
considers the synthesis of 3, its O,O-dimethyl analogue
(7), and related compounds (Figure 1) and their biological
properties and degradation chemistry. The findings are
of special interest since we observe that 3 undergoes
alternative reactions proposed to involve N f O rear-
rangement or fragmentation through a metaphosphate
analogue.
Exp er im en ta l P r oced u r es
Toxicity in Mice a n d Hou seflies. Male albino Swiss-
Webster mice (21-23 g) from Simonsen Laboratories (Gilroy,
CA) were treated ip with the test compounds using, as carrier
vehicles (50 µL), water for 1 and 2, aqueous Me2SO for 3 and 7,
and Me2SO alone for all other toxicants and oxidase inhibitors.
Adult female houseflies (Musca domestica L., SCR strain
cultured in this laboratory, ∼20 mg each, maintained at 25 °C)
were treated with the test compounds using acetone (0.2 µL)
applied as a measured drop to the ventrum of the abdomen.
The severity of poisoning signs for mice and the knockdown (KD)
of flies (inability to walk or fly) were rated at 30 and 150 min,
respectively, and LD50 values were determined at 24 h. Three
oxidase inhibitors [piperonyl butoxide (PB) (Fluka Chemical
Corp., Ronkonkoma, NY) and methimazole and N-benzylimi-
dazole (NBI) (Aldrich Chemical Co., Milwaukee, WI)] were
tested for possible effects in delaying the toxicity of 1 and
parathion in mice and houseflies.
Sp ectr oscop y, Ch r om a togr a p h y, a n d An a lysis. 1H and
1
31P NMR spectra (the latter without H coupling) were recorded
with a Bruker WM-300 spectrometer at 300 and 121.5 MHz,
respectively. Chemical shifts are referenced to internal 3-(tri-
methylsilyl)propionic-2,2,3,3-d4 acid, sodium salt, or tetrameth-
ylsilane for 1H in D2O and CDCl3, respectively, and to external
phosphoric acid in the appropriate solvent for 31P. GC/MS/
chemical ionization (CI) involved Hewlett-Packard 5890 gas
chromatograph coupled to a 5971A mass spectrometer; DB-5
fused silica gel capillary column, 30 m × 0.25 mm i.d. (J and W
Scientific, Folsom, CA); injection port temperature 250 °C;
temperature program 70-250 °C over 18 min; tR 9.9 min for
2-oxohexamethylenimine (Aldrich Chemical Co., Milwaukee,
WI) and 4.8 min for (MeO)2P(O)(OMe). Fast atom bombard-
ment/high-resolution mass spectrometry (FAB/HRMS) was
performed with a Kratos MS-50 instrument (Department of
Br a in ACh E In h ibition in Mice a n d Hou seflies. Mouse
brain and housefly heads were homogenized at 1% and 3% (w/
v), respectively, in 100 mM phosphate buffer (pH 7.4), and the
housefly preparation was clarified by passing through cheese-
2
N. J . Cherrington, J . G. Falls, R. L. Rose, K. M. Clements, R. M.
Philpot, P. E. Levi, and E. Hodgson, unpublished results.

28 Chem. Res. Toxicol., Vol. 11, No. 1, 1998
Sch em e 2. Syn th esis Rou tes for th e P r in cip a l Com p ou n d s
Mahajna and Casida
Chemistry, University of California, Berkeley). Elemental
analyses were determined by the Microchemical Analysis
Laboratory of the Chemistry Department at Berkeley.
Ca u tion : Some of the organophosphorus compounds consid-
ered are known to be toxic, and they should all be studied under
containment conditions.
Ch em ica ls. The insecticides examined (>99% purity) were
1 from Chevron Chemical Co. (Richmond, CA) and parathion
from Chem Service (West Chester, PA). Other organophospho-
rus compounds studied were from syntheses described below.
Syn th esis. Gen er a l. The compounds considered are phos-
phoramidates (1 and 2), N-hydroxyphosphoramidates and de-
rivatives thereof (3-10), and O-amino phosphates (11 and 12
as proposed intermediates and 13-15 as related chemicals)
(Figure 1). Synthesis routes for the principal compounds are
given in Scheme 2. All compounds prepared were obtained in
high purity (>98%) based on 1H NMR and FAB/HRMS as
specified plus elemental analyses in the case of 7 nosylate and
10 nosylate.
Syn t h esis of N-H yd r oxyp h osp h or a m id ot h iola t es a n d
Der iva tives 3-6. In preparing 3, the preferred reagent was
Me3SiNHOSiMe3 rather than H2NOSiMe3 or H2NOH in agree-
ment with earlier studies (22-24). MeO(MeS)P(O)Cl (16) was
allowed to react with Me3SiNHOSiMe3 in CH2Cl2 to give MeO-
(MeS)P(O)NHOSiMe3 (17), requiring pyridine catalysis, and
desilylation with MeOH yielded 3. Similarly, 4 and 5 were
obtained by the reaction of 16 with H2NOMe and H2NOBn,
in dry CH2Cl2 (30 mL) at -30 °C under N2. After stirring for 4
h at ambient temperature, the solvents were evaporated, dry
ether was added, pyridinium chloride was removed by filtration,
and the residue from evaporation of the ether was chromato-
graphed on silica gel. Compound 4 (colorless oil) was eluted
with 5% EtOAc/hexane, yield 76%. NMR (CDCl3): 1H δ 2.36
(3H, d, J ) 15 Hz), 3.72 (3H, s), 3.86 (3H, d, J ) 12 Hz); 31P δ
39.6. FAB/HRMS: [MH+] calcd for C3H11NO3PS, 172.0197;
found, 172.0196.
MeO(MeS)P (O)NHOBn (5): colorless oil, prepared using
NH2OBn hydrochloride and purified as described for 4. NMR
(CDCl3): 1H δ 2.37 (3H, d, J ) 15 Hz), 3.88 (3H, d, J ) 11 Hz),
4.84 (2H, broad), 7.35 (5H, m); 31P δ 40.23.
MeO(MeS)P (O)NMeOH (6): colorless oil, prepared by treat-
ing 16 with MeNHOSiMe3, following the same procedure used
to obtain 3. NMR (CDCl3): 1H δ 2.37 (3H, d, J ) 15 Hz), 3.10
(3H, d, J ) 12 Hz), 3.81 (3H, d, J ) 12 Hz); 31P δ 42.8.
Syn th esis of N-Hyd r oxyp h osp h or a m id a tes 7-10. Simi-
lar procedures to those above but from (MeO)2P(O)Cl (18) gave
N-hydroxyphosphoramidates 7-10. While O-trimethylsilyl de-
rivatives 19 and 20 (see below) are stable for hours at 25 °C
and weeks at -20 °C (under N2), 7 and 10 decompose slowly at
1
room temperature and were characterized by H and 31P NMR
and, in addition, by derivatization with nitrobenzenesulfonyl
chloride (NsCl) and Et3N (24) to obtain the crystalline p-
nitrobenzenesulfonyl derivatives (nosylates), i.e., 7 nosylate and
10 nosylate. The O-substituted derivatives 8 and 9 were stable
as neat compounds at room temperature and purified by column
chromatography.
respectively, using pyridine to scavenge the liberated HCl.
6
was prepared by treating 16 with MeNHOSiMe3 (25) followed
by desilylation with MeOH. Compounds 3-6 were stable under
N2 for several hours at room temperature and several weeks at
-20 °C.
(MeO)2P (O)NHOSiMe3 (19). Compound 18 (Aldrich) was
treated with Me3SiNHOSiMe3, as in the preparation of 3 but
without desilylation, to yield 19 as an oil (crude yield 82%). NMR
(CDCl3): 1H δ 0.15 (9H, s), 3.78 (6H, d, J ) 11 Hz); 31P δ 12.11.
(MeO)2P (O)NHOH (7): obtained as an oil on treatment of
19 with anhydrous MeOH for 5 min and solvent evaporation.
NMR (CDCl3): 1H δ 3.81 (6H, d, J ) 11 Hz); 31P δ 13.15. For
further characterization as the nosylate, 7 (3 mmol) was
dissolved in ice-cold CH2Cl2 (5 mL), and NsCl (3.6 mmol) was
immediately added followed by Et3N (3.6 mmol). The resulting
reaction mixture was stirred for 6 h and then kept at -20 °C
overnight. 7 nosylate was filtered off, washed thoroughly with
ether, and recrystallized from cold aqueous methanol (yield
74%), mp 128-129 °C dec. NMR (CDCl3/CD3OD, 3:1): 1H δ 3.72
(6H, d, J ) 11 Hz), 8.23 (2H, d, J ) 9 Hz), 8.48 (2H, d, J ) 9
Hz); 31P δ 11.71. Anal. Calcd for C8H10N2O8PS: C, 29.55; H,
3.10; N, 8.61; S, 9.86. Found: C, 29.80; H, 3.47; N, 8.60; S, 9.93.
(MeO)2P (O)NMeOSiMe3 (20). A procedure identical with
that used to prepare 19 but from MeNHOSiMe3 was followed.
NMR (CDCl3): 1H δ 0.17 (9H, s), 3.11 (3H, d, J ) 12 Hz), 3.68
(6H, d, J ) 11 Hz); 31P δ 8.63.
MeO(MeS)P (O)Cl (16). In a procedure involving chlorina-
tion and thermal isomerization, a mixture of (MeO)2P(S)OMe
(Alfa, Ward Hill, MA) (10 mmol) and POCl3 (10 mmol) was
stirred at 60-70 °C for 8 h and then distilled to give 16 as a
colorless oil (72%), bp 57-58 °C/1 mmHg. NMR (CDCl3): 1H δ
2.49 (3H, d, J ) 19 Hz), 3.94 (3H, d, J ) 15 Hz); 31P δ 39.8.
MeO(MeS)P (O)NHOH (3). To a solution of 16 (30 mmol)
in dry CH2Cl2 (30 mL), stirred at 0 °C under N2, was added
dropwise a solution of Me3SiNHOSiMe3 (Aldrich) (30 mmol) and
pyridine (2 mL) in dry CH2Cl2 (30 mL). After stirring for 6 h
at room temperature and solvent removal at reduced pressure,
the residue was taken up in anhydrous ether, which was filtered
and evaporated to dryness. Dry CH2Cl2 (30 mL) was added
followed by absolute MeOH (10 mL) with stirring for 5 min and
then evaporation to give 3 (31%) as an oil that crystallized from
CH2Cl2 after a few days at -20 °C, mp 75-77 °C dec. NMR
(CDCl3): 1H δ 2.33 (3H, d, J ) 15 Hz), 3.85 (3H, d, J ) 12 Hz);
31P δ 44.2. FAB/HRMS: [MH+] calcd for C2H9NO3PS, 158.0040;
found, 158.0044.
(MeO)2P (O)NMeOH (10): obtained as an oil from 20 on
methanolysis as for 7. NMR (CDCl3): 1H δ 3.09 (3H, d, J ) 12
Hz), 3.72 (6H, d, J ) 11 Hz); 31P δ 9.13. For further charac-
terization, 10 nosylate was obtained as described above as a
MeO(MeS)P (O)NHOMe (4). A solution of H2NOMe hydro-
chloride (30 mmol) in dry CH2Cl2 (30 mL) was added dropwise
with stirring to a solution of 16 (30 mmol) and pyridine (20 mL)

N-Hydroxymethamidophos Chemistry and Action
Chem. Res. Toxicol., Vol. 11, No. 1, 1998 29
hygroscopic solid. NMR (CDCl3): 3.06 (3H, d, J ) 12 Hz), 3.65
(6H, d, J ) 11 Hz), 8.17 (2H, d, J ) 9 Hz), 8.39 (2H, d, J ) 9
Hz); 31P δ 8.91. Anal. Calcd for C9H13N2O8PS: C, 31.77; H,
3.85; N, 8.23; S, 9.42. Found: C, 32.10; H, 3.99; N, 7.97; S, 9.05.
(MeO)2P (O)NHOMe (8). The procedure used to prepare 4
but from 18 was followed to obtain 8 as an oil. NMR (CDCl3):
1H δ 3.67 (3H, s), 3.84 (6H, d, J ) 11 Hz); 31P δ 10.54. FAB/
HRMS: [MH+] calcd for C3H11NO4P, 156.0425; found, 156.0427.
(MeO)2P (O)NHOBn (9): oil, prepared and purified as de-
scribed for 5 but from 18. NMR (CDCl3): 1H δ 3.89 (6H, d, J )
11 Hz), 4.78 (2H, broad), 7.36 (5H, m); 31P δ 9.71.
Syn th esis of O-Am in o P h osp h a te Der iva tives 13-15. To
prepare 13, MeN(OH)CO2-t-Bu (26) was treated with 18 to give
N-protected 13 (i.e., 21) followed by deprotection on treatment
with trifluoroacetic acid. Compound 13 was an oil that decom-
posed gradually at room temperature (stable at -20 °C for
several days). 14 was obtained by reaction of 16 with Me2NOH
in the presence of pyridine.
(MeO)2P (O)ON(Me)CO2-t-Bu (21): a stable oil, synthesized
by allowing MeN(OH)CO2-t-Bu to react with 18, following the
procedure used to prepare 4 (yield 79%). NMR (CDCl3): 1H δ
1.48 (9H, s), 3.20 (3H, s), 3.88 (6H, d, J ) 11 Hz); 31P δ 2.86.
(MeO)2P (O)ONHMe (13): obtained on dropwise addition of
trifluoroacetic acid (1.5 mL) to an ice-cooled solution of 21 (1
mmol) in CH2Cl2 (5 mL), stirring the mixture for 5 h at 0 °C,
and then pouring into ice-water and immediately extracting
with CH2Cl2. The extract was dried (Na2SO4) and concentrated
(colorless oil) (yield 43%). NMR (CDCl3): 1H δ 3.11 (3H, s), 3.86
(6H, d, J ) 11 Hz); 31P δ 3.28. FAB/HRMS: [MH+] calcd for
C3H11NO4P, 156.0425; found, 156.0418.
F igu r e 2. Effect of methimazole and N-benzylimidazole on
brain acetylcholinesterase inhibition from methamidophos in
mice. The oxidase inhibitors were administered ip at 100 mg/
kg 30 min before ip treatment with 1 at 1.25, 2.5, 5, or 7.5 mg/
kg and determination of brain AChE activity 30 min later.
Ta ble 1. Effect of Th r ee Oxid a se In h ibitor s on Br a in
Acetylch olin ester a se In h ibition a n d Toxicity fr om
Meth a m id op h os a n d P a r a th ion in Mice a n d Hou seflies
methamidophosb
parathionb
AChE inhibn,
AChE inhibn,
pretreatmenta
%
toxicity
Mice
%
toxicity
none
PB
NBI
methimazole
86 ( 2
81 ( 3
38 ( 3
11 ( 2
++++
92 ( 3
87 ( 5
18 ( 4
99 ( 1
++++
+++
+
++++
++
(MeO)2P (O)ONMe2 (14): prepared as an oil by reaction of
18 and HONMe2 hydrochloride in dry ether/pyridine at -20 °C,
following a similar procedure to that used for preparation of 4.
NMR (CDCl3): 1H δ 2.82 (6H, s), 3.83 (6H, d, J ) 11 Hz); 31P δ
3.76. FAB/HRMS: [MH+] calcd for C4H13NO4P, 170.0582;
found, 170.0579.
+
++++
Houseflies
none
PB
NBI
methimazole
100
100
13 ( 2
7 ( 1
99
99
14
1
65 ( 3
49 ( 2
4 ( 1
96
66
1
47 ( 3
71
(MeO)2P (O)ONdC6H10 (15): prepared by treating 18 with
cyclohexanone oxime, following the procedure used to prepare
4, and purified by preparative TLC on silica gel, Rf ) 0.25 in
EtOAc/hexane, 3:2. NMR (CDCl3): 1H δ 1.65 (6H, m), 2.27 (2H,
t), 2.52 (2H, t), 3.81 (6H, d, J ) 11 Hz); 31P δ 3.32. FAB/
HRMS: [MH+] calcd for C8H17NO4P, 222.0895; found, 222.0900.
Hyd r olysis. The hydrolysis of compounds 1-10 and 13-
15 (250 mM) in the appropriate medium (including ∼10% D2O
as the locking solvent) was carried out at room temperature
(unless indicated otherwise) monitoring the progress by inte-
grating the signal intensities for the starting material versus
the product(s) by 31P NMR spectroscopy. Values for t1/2 were
derived from at least five points in the range of 10-90%
hydrolysis. In a few cases the aqueous solution was then
extracted with CH2Cl2 which was dried over Na2SO4 and
analyzed by GC/MS/CI.
a
Mice were treated ip with each oxidase inhibitor at 100 mg/
kg and 30 min later with 1 at 12 mg/kg or parathion at 14 mg/kg.
Houseflies were treated topically with PB at 100 µg/g or NBI or
methimazole at 500 µg/g followed after 30 min by 1 at 15 µg/g or
b
parathion at 3 µg/g. AChE inhibition (mean ( standard error, n
) 3) and toxicity in mice (poisoning signs rated from + for mild to
++++ for most severe, n ) 6) and houseflies (KD, %, n ) 150)
were determined 30 and 150 min after toxicant treatment in mice
and houseflies, respectively. The oxidase inhibitors alone gave no
AChE inhibition or poisoning signs.
i.e., methimazole is more effective in protecting against
1 and NBI against parathion (Table 1). Although meth-
imazole and NBI greatly reduce the rate of KD in
houseflies, they have little or no effect on the 24-h LD50
values (data not given). These results are interpreted
as 1 undergoing bioactivation by methimazole- and NBI-
sensitive oxidases in both mice and houseflies.
Resu lts
P ossible Bioa ctiva tion of 1 a s a n ACh E In h ibitor
a n d Meta bolism in in Vitr o System s. The inhibitory
potency of 1 for mouse brain AChE activity is not changed
under the following incubation conditions: NADPH
alone, mouse liver microsomes, or FMO 1; addition of
NADPH to either oxidase preparation. Under these
assay conditions, 1 is not activated as an inhibitor of
brain AChE by NADPH-dependent systems in mouse
brain homogenate or mouse liver microsomes or by a
combination of NADPH and FMO 1.
[14CH3S]1 is a poor substrate or not a substrate for the
oxidase systems examined. Incubation with mouse liver
microsomes results in up to 10% substrate loss mostly
as polar metabolite(s) appearing at the solvent front on
HPLC; this apparent metabolism is independent of
NADPH fortification. Substrate recovery is complete
Effect of Oxid a se In h ibitor s on Br a in ACh E
In h ibition a n d Toxicity fr om 1 a n d P a r a th ion in
Mice a n d Hou seflies. Mouse brain AChE activity is
progressively inhibited by 1 at 1.25-7.5 mg/kg 30 min
after treatment, and this inhibition is partially blocked
by pretreatment with NBI and almost completely with
methimazole (Figure 2). In a second study with 1 at 12
mg/kg in mice, methimazole is again more effective than
NBI, and PB is essentially inactive as a protectant for
both AChE inhibition and poisoning signs (Table 1).
Parathion (as a positive control) follows a similar rela-
tionship in mice except that, with both AChE inhibition
and poisoning signs, the order for oxidase inhibitor
effectiveness is NBI . PB or methimazole (Table 1).
Housefly head AChE inhibition and KD of the flies by 1
and parathion follow the same pattern as with the mice,

30 Chem. Res. Toxicol., Vol. 11, No. 1, 1998
Mahajna and Casida
Ta ble 2. Toxicity to Mice a n d Hou seflies a n d
An tich olin ester a se Activity of O,S-Dim eth yl a n d
O,O-Dim eth yl P h osp h or a m id a tes
Ta ble 3. Effect of p H on th e Ra te of Hyd r olysis a t Room
Tem p er a tu r e of (MeO)2P (O)NHOH (7) a n d
MeO(MeS)P (O)NHOH (3) a n d on th e Ra tio of P r od u cts
fr om 3
mouse brain AChE
b
inhibition, IC50
MeO(MeS)P(O)NHOH
micea housefliesa
R
LD50
,
LD50
µg/g
,
in vitro,
µM
in vivo,
mg/kg
(MeO)2P(O)NHOH
MeOP(O)(OH)2 ÷
pHa
t
1/2, min
t1/2, min
MeO(MeS)P(O)OHb
no. substituent mg/kg
MeO(MeS)P(O)R
1.2
1.8
3.0
5.1
7.4
8.6
1
4
27
80
110
47
33
9
0.17
0.31
6.0
0.75
0.67
2.5
1
3
4
5
NH2
11
40
>250
115
57
4
NHOH
NHOMe
NHOBn
5
>500
6
>1000 (0%)c,d ∼60 (41%)c
220
10
28
40
21
9.9
10.9
(MeO)2P(O)R
5
1.2
2
7
NH2
NHOH
>250
>500
240
>150 (0%)c
>250e >500e
>1000 (0%)c,d >250 (0%)c
a
Buffers (100 mM, pH) were glycine (1.8), succinate (3.0, 5.1),
Tris-HCl (7.4, 8.6), and Na carbonate (9.9, 10.9). Buffer catalysis
a
ip for mice and topical for houseflies. LD50 values determined
by log-probit analyses using composite data from three experi-
ments. Assay conditions described in Experimental Procedures.
b
might be a major contributor to the pH-rate profile. Product
b
ratio determined by relative 31P NMR signal intensity.
IC50 values (concentration for 50% inhibition) based on three
slower at pH 3.0 and 5.1; the major products are MeO-
(MeS)P(O)OH and MeOP(O)(OH)2 under acidic and basic
conditions, respectively (Table 3). A minor and transient
31P NMR signal at 36.2 ppm in aqueous solution disap-
pears immediately on adding cyclohexanone (100 µL), and
2-oxohexamethylenimine is apparent by GC/MS in the
organic extract of the reaction mixture (MH+ 114); these
products are considered later on analogy with studies on
the hydrolysis of 7 (Figure 3).
experiments. In vivo data for 1 from Figure 2. c Percent inhibition
d
at indicated concentration or dose. No inhibition observed at 1000
µM with either bovine erythrocyte AChE or chymotrypsin on 30
min preincubation before substrate addition. e LD50 values for 8
and 9: >250 mg/kg for mice and >500 µg/g for houseflies.
following incubation of [14CH3S]1 with FMO 1 in the
presence or absence of NADPH. Finally, the rate of
NADPH oxidation by FMO 1 is enhanced by methimazole
but not by 1 (data not given).
The O-methyl and O-benzyl derivatives of 3 (i.e., 4 and
5) are stable under the conditions described above. In
contrast, the N-methyl derivative 6 decomposes at a
moderate rate (t1/2 ) 150 min, room temperature) to give
an intermediate 31P NMR (D2O) signal at δ 40.5 and
ultimately MeO(MeS)P(O)OH as the only phosphorus-
containing product. This intermediate from 6 is more
stable than the corresponding one at δ 36.2 from 3.
Hyd r olysis of (MeO)2P (O)NHOH (7) a n d Der iva -
tives 8-10. Compound 7 decomposes rapidly (t1/2 ) 15
min, room temperature) in H2O to yield (MeO)2P(O)OH
and in H218O to give (MeO)2P(O)18OH [determined as
(MeO)2P(O)18OMe after methylation with diazomethane
(in ether/THF) by GC/MS, MH+ 143]. Hydrolysis is most
rapid in acidic conditions and slowest at pH 5.1 with t1/2
) 10 min at pH 7.4 (Table 3). Decomposition of 7 in
MeOH, EtOH, or various alcohol-water mixtures always
gives (MeO)2P(O)OH as the only phosphorus-containing
product, although the reaction rate is depressed strongly
by alcohol, e.g., t1/2 in MeOH > 48 h. 31P NMR examina-
tion of the reaction of 7 at pH 5-11 reveals an interme-
diate signal at 5.4 ppm which is more intense at pH 9-11
or when alcohol is present in the solvent mixture. In
carbonate buffer (pH 11)-methanol (3:1) there is rapid
disappearance of the starting material (<5 min) and
appearance of this 5.4 ppm signal which is then replaced
by a 4.6 ppm signal on addition of cyclohexanone to the
reaction mixture. The 4.6 ppm signal, assigned as
(MeO)2P(O)ONdC6H10 (15) (with confirmation by signal
enhancement on adding the authentic standard), then
decays (<10 min) and is replaced by a signal correspond-
ing to (MeO)2P(O)OH (Figure 3). The reaction mixture
also contains 2-oxohexamethylenimine identified by GC/
MS. In a separate experiment 15 was shown to undergo
rapid decomposition in water (t1/2 ) 5 min, room tem-
perature) yielding (MeO)2P(O)OH and oxohexamethyl-
enimine, most likely via a Beckmann-like rearrangement
(Figure 3). The tendency of 15 to undergo rapid rear-
rangement may be associated with (MeO)2P(O)O- serving
as a good leaving group.
Toxicity to Mice a n d Hou seflies a n d An tich o-
lin est er a se Act ivit y of O,S-Dim et h yl a n d O,O-Di-
m eth yl P h osp h or a m id a tes (Ta ble 2). LD50 values for
N-hydroxymethamidophos (3) and its O-benzyl ether (5)
are 4-10-fold greater than for methamidophos (1) in both
mice and houseflies, whereas the O-methyl derivative (4)
is essentially inactive. The O,O-dimethyl phosphorami-
dates 2 and 7-9 are not toxic under the test conditions.
AChE inhibitory activity was directly compared with
mouse brain enzyme for phosphoramidates 1 and 2 and
N-hydroxyphosphoramidates 3 and 7. Methamidophos
(1) is 4-fold more potent than its O,O-dimethyl analogue
(2) in vitro and >38-fold in vivo consistent with in vivo
bioactivation of 1. The N-hydroxy compounds are es-
sentially inactive in vitro (even at 1000 µM) and in vivo
(except at a lethal dose for 3 with mice showing typical
signs of cholinergic poisoning). In other studies 3 and 7
are also inactive at 1000 µM as inhibitors of bovine
erythrocyte AChE and chymotrypsin.
Hyd r olysis a n d F r a gm en ta tion of MeO(MeS)P -
(O)NHOH (3) a n d Der iva tives 4-6. Compound 3
hydrolyzes rapidly (t1/2 ) 10 min, 37 °C) at pH 7.4 to give
O,S-diester MeO(MeS)P(O)OH [δ 24.23 (D2O)] (for syn-
thesis, see ref 19) and monoester MeOP(O)(OH)2 [δ 1.67
(D2O)] in a 3:2 ratio, respectively. This ratio of products
is constant throughout the reaction. In a control experi-
ment, the diester gives no detectable monoester after 48
h. While 3 hydrolyzes quickly in water and D2O (t1/2
)
56 and 60 min, respectively, room temperature) giving
MeO(MeS)P(O)OH and MeOP(O)(OH)2, it reacts slowly
in absolute MeOH and EtOH (t1/2 ) 55 and 68 h,
respectively) to give a single product in each case,
(MeO)2P(O)OH [δ 3.10 (D2O)] and MeO(EtO)P(O)OH [δ
2.14 (D2O)], respectively. In 50% aqueous EtOH, the
products from 3 are MeO(MeS)P(O)OH, MeO(EtO)P(O)-
OH, and MeOP(O)(OH)2 in a molar ratio of 4:3:3. The
hydrolysis rate and product ratio at room temperature
in buffered solutions depend on the pH; fragmentation
is rapid under basic and strong acidic conditions but

N-Hydroxymethamidophos Chemistry and Action
Chem. Res. Toxicol., Vol. 11, No. 1, 1998 31
F igu r e 3. Mechanism for formation of 2-oxohexamethylenimine on reaction of O,O-dimethyl N-hydroxyphosphoramidate (7) with
cyclohexanone in aqueous medium. A partial reaction sequence is also shown for 3 under the same conditions.
The hydrolysis mechanism of 7 was also examined with
analogues in which the hydroxyl oxygen was blocked with
methyl (8) or benzyl (9) or nitrogen was substituted with
methyl (10). No fragmentation is observed for the
O-substituted derivatives 8 and 9 in water for 24 h, but
in contrast 10 in water at ambient temperature under-
microsomes and mitochondria (presumably P450) acti-
vate parathion in vitro as an AChE inhibitor (29) but
again apparently not 1 from the present studies with
brain homogenate. Although 1 is not bioactivated in vitro
as an AChE inhibitor or metabolized in an NADPH-
dependent manner by mouse liver microsomes and FMO
1, these coupled activation-target system assays may not
detect a very unstable intermediate.
The unusually facile peracid oxidation of 1 (5, 7, 30)
involves mostly attack at sulfur but 1-SO and the
corresponding sulfone are not observed, presumably due
to instability (11). In contrast to the chemical model
systems in organic solvents, 1-SO may be more stable in
an appropriate aqueous enzymatic system if sulfoxidation
is more facile than further oxidation or cleavage reac-
tions. In any case, 1-SO has not been synthesized and
remains a candidate, but not established, bioactivation
product.
N-Hydroxylation was considered as an alternative to
S-oxidation for bioactivation of 1. N-Hydroxymethami-
dophos (3) was prepared to study its biological activity
and chemical reactivity as a basis for judging its rel-
evance as a candidate bioactivation product. The toxic
action of 3, although less potent than 1, is attributable
to cholinergic poisoning in mice, yet 3 is essentially
inactive in vitro as an AChE inhibitor. This may be due
to differences in the assay conditions relative to the
instability of 3 tested directly or formed as a metabolite
of 1. Accordingly, it is important to understand the
hydrolysis and fragmentation of 3 in aqueous systems,
and to this end a comparison was made with the nontoxic
O,O-dimethyl analogue 7.
The hydrolysis and fragmentation of 3 and 7 are
proposed to involve two novel mechanisms: N f O
rearrangement for both compounds (mechanism A) and
fragmentation via a metaphosphate analogue for 3 only
(mechanism B) (Figure 4). Mechanism A results in
cleavage of the P-N bond and formation of MeO(MeS)P-
(O)OH from 3 and (MeO)2P(O)OH from 7. This is not
direct hydrolysis because: (a) hydroxylamine as a strong
base is a poor leaving group, exemplified by the relative
stability of hydroxamic acids; (b) displacement of hy-
droxylamine by water via an “SN2-like” mechanism
requires a large phosphonium ion character with an
extended P-N bond breaking in the transition state. N
f O rearrangement, involving nucleophilic attack of the
hydroxylamine oxygen at phosphorus, is therefore pro-
posed as the first step followed by nucleophilic displace-
ment of the ONH2 by OH (from H2O) to give MeO(MeS)P-
goes similar fragmentation to that of 7 but slower (t1/2
)
180 min). Most importantly, O-amino phosphate deriva-
tive 13 is assigned as an intermediate in the hydrolysis
of 10 based on 31P NMR [δ 4.92 (D2O)] and confirmed by
matching with an authentic standard.
H yd r olysis of O-Am in o P h osp h a t e Der iva t ives
12-14. Compound 12, as an intermediate in the frag-
mentation of 7, decomposes rapidly in water at room
temperature (t1/2 < 5 min) to give (MeO)2P(O)OH. Under
similar conditions, (MeO)2P(O)ONHMe (13) hydrolyzes
>20 times slower than 12 and in [18O]water gives
(MeO)2P(O)18OH as the only phosphorus-containing prod-
uct. (MeO)2P(O)ONMe2 (14) shows relatively high stabil-
ity, and only 37% hydrolysis to (MeO)2P(O)OH is ob-
served after 100 h in water at room temperature.
Discu ssion
Methamidophos is firmly established for the first time
in this investigation to undergo oxidative bioactivation
as an AChE inhibitor and toxicant in both mammals and
insects. The most compelling evidence is our finding that
oxidase inhibitors delay the AChE inhibition and toxicity
from 1 in mice and houseflies as observed earlier for
parathion [the classical bioactivated AChE inhibitor (27)].
The most potent compounds, methimazole and NBI, delay
the action of 1 in houseflies but ultimately have little or
no effect on the LD50 value; this may be due to more rapid
metabolism of the oxidase inhibitors than of 1 leading to
recovery of the oxidase activity and continued activation
of the insecticide. On analogy with the parathion-
paraoxon relationship (27), bioactivated 1 is probably
inhibitory for AChE at 1-100 nM.
The enzyme(s) responsible for bioactivation of 1 in mice
and houseflies is not established. NADPH-dependent
and PB-sensitive P450s in mouse liver microsomes
activate the phosphorothionate parathion (27) and many
S-ethyl, S-propyl, and S-butyl phosphorothiolates but not
the S-methyl phosphorothiolate 1 (ref 6 and this study).
The different specificities of oxidase inhibitors in delaying
poisoning by 1 and parathion, favoring methimazole for
1 and NBI for parathion, suggest that activation may
involve FMOs for 1 and P450s for parathion (28). Brain

32 Chem. Res. Toxicol., Vol. 11, No. 1, 1998
Mahajna and Casida
F igu r e 4. Proposed N f O rearrangement of O,S-dimethyl and O,O-dimethyl N-hydroxyphosphoramidates (3 and 7, respectively)
to the corresponding O-amino phosphates (11 and 12, respectively) (mechanism A) and fragmentation of 3 via the phosphimino
tautomer and a metaphosphate analogue (metaphosphorimidate) (mechanism B).
(O)OH or (MeO)2P(O)OH. In a related example from
carbonyl chemistry, interestingly the O-acylhydroxyl-
amines undergo O f N rearrangement to hydroxamic
acids, i.e., RC(O)ONH2 f RC(O)NHOH (31). The inter-
mediacy of the O-amino phosphates, i.e., N f O rear-
rangement, is established in four ways: observation of
an intermediate with a 31P NMR chemical shift 7-10
ppm upfield relative to the PNH2 compound with the
proposed MeO(MeS)P(O)ONH2 less stable than (MeO)2P-
(O)ONH2; direct observation of 13 (verified by synthesis)
as an intermediate in the hydrolysis of 10; blocking the
hydroxyl substituent of 3 and 7 as in 4, 5, 8, and 9 giving
stable derivatives by preventing the rearrangement
reaction; and identification, on adding cyclohexanone, of
15 from 7 in water and 2-oxohexamethylenimine from
both 3 and 7 in water (Figure 3) [this is analogous to the
formation of O-acyl oximes on addition of cyclohexanone
as structural proof for O-acylhydroxylamines (32)].
cleophiles of sufficiently similar solvation properties (34),
which proved to be the case with 3 in aqueous EtOH.
Thus, the rate-limiting step is a unimolecular process in
fragmentation to the metaphosphate analogue and then
nonselective reaction with the solvents. N-Hydroxymetha-
midophos (3) is expected to ionize largely in the phos-
phimino form [as for the carbonyl analogues, the hydrox-
amic acids, in the oximino form (35)] conferring high
reactivity caused by the lone electron pairs of the
adjacent oxygen atom, i.e. R-effect (36). The unstable
phosphimino tautomer loses MeSH to form the meta-
phosphate analogue that is trapped by water or an
alcohol to give the demethylthio derivative. The initially
formed N-hydroxyphosphorimidate then rearranges, and
the O-amino phosphate undergoes hydrolysis. In acidic
conditions 3 hydrolyzes mainly via N f O rearrangement
(mechanism A), while in basic conditions mechanism B
is favored and MeOP(O)(OH)2 is obtained as the major
product; once the hydroxylamine oxygen of 3 is negatively
charged (basic pH), the R-effect is stronger and the
movement of the electron pair from nitrogen into the
P-N bond is made easier.
Metaphosphate analogue pathway B in fragmentation
of 3 requires tautomerization of the NH substituent to
the phosphimino form and therefore should not take place
with the N-methyl derivative [e.g., the analogous N-
substituted hydroxamic acid cannot tautomerize to the
oximino form and instead ionizes in the normal way (35)].
This proved to be the case for 6 which has only the N f
O rearrangement pathway (A) yielding MeO(MeS)P(O)-
OH but not MeOP(O)(OH)2. The greater stability of
O-substituted analogues 4 and 5 under similar condi-
tions, even though in terms of the R-effect they are
expected to hydrolyze via mechanism B to give MeOP-
The novel PNHOH f PONH2 rearrangement is highly
dependent on both the solvent and pH. Compounds 3
and 7 are stable in CHCl3 and Me2SO but rearrange
slowly in alcohol and rapidly in water. The facile
hydrolysis in water is attributable to stabilization of the
transition state and better solvation of the leaving group
via hydrogen bonds. The reaction is catalyzed by both
acid and base with a minimum rate in the region of
hydroxylamine dissociation (pKa 6.0) (33).
Dissociative fragmentation is proposed as a second
mechanism (B) to explain the formation, in addition to
MeO(MeS)P(O)OH via mechanism A, of MeOP(O)(OH)2
in water and (MeO)2P(O)OH and MeO(EtO)P(O)OH in
MeOH and EtOH, respectively (Figure 4). A metaphos-
phate analogue (metaphosphorimidate) is proposed as the
intermediate to explain the products formed in the
fragmentation of 3.3 A diagnostic test for involvement
of a metaphosphate as an intermediate is the absence of
a steric effect in the phosphorylation of competing nu-
3
Earlier speculations on a metaphosphate analogue intermediate
in the base-catalyzed hydrolysis of 1 (2, 3) and the transitory nature
of 1 sulfoxide and its enzyme inhibitory potential (8) were not
supported by any experimental evidence.

N-Hydroxymethamidophos Chemistry and Action
Chem. Res. Toxicol., Vol. 11, No. 1, 1998 33
(5) Eto, M., Okabe, S., Ozoe, Y., and Maekawa, K. (1977) Oxidative
activation of O,S-dimethyl phosphoramidothiolate. Pestic. Bio-
chem. Physiol. 7, 367-377.
(6) Wing, K. D., Glickman, A. H., and Casida, J . E. (1983) Oxidative
bioactivation of S-alkyl phosphorothiolate pesticides: Stereospeci-
ficity of profenofos insecticide activation. Science 219, 63-65.
(7) Bellet, E. M., and Casida, J . E. (1974) Products of peracid
oxidation of organothiophosphorus compounds. J . Agric. Food
Chem. 22, 207-211.
(8) Thompson, C. M., Castellino, S., and Fukuto, T. R. (1984) A
carbon-13 nuclear magnetic resonance study on an organophos-
phate. Formation and characterization of methamidophos (O,S-
dimethyl phosphoramidothioate) S-oxide. J . Org. Chem. 49, 1696-
1699.
(O)(OH)2, is resolved below by studying the behavior of
the O-amino phosphates.
O,O-Dimethyl O-amino phosphate and its N-methyl
derivatives (12-14) hydrolyze by displacement of the
ONR2 group with OH from water at a rate progressively
reduced on replacing the hydrogens with methyl groups
(i.e., 12 > 13 . 14). Similarly in the O,S-dimethyl series,
the NHMe intermediate from 6 [MeO(MeS)P(O)ONHMe]
is more stable than the NH2 intermediate from 3 (i.e.,
11). As analogues of the organic peracids (a ) (37) and
the O-acylhydroxylamines (b), the O-amino phosphates
(c) are expected to be intramolecularly hydrogen-bonded
with accompanying enhancement of the inductive effect
of the phosphoryl group (38).
(9) Suksayretrup, P., and Plapp, F. W., J r. (1977) Mechanisms by
which methamidophos and acephate circumvent resistance to
organophosphate insecticides in the housefly. J . Agric. Food
Chem. 25, 481-485.
(10) Kao, T.-S., and Fukuto, T. R. (1977) Metabolism of O,S-dimethyl
propionyl- and hexanoylphosphoramidothioate in the house fly
and white mouse. Pestic. Biochem. Physiol. 7, 83-95.
(11) Wu, S.-Y., Toia, R. F., and Casida, J . E. (1992) Phosphorothiolate
sulfoxides and sulfones: NMR characteristics and reactivity. J .
Agric. Food Chem. 40, 1425-1431.
(12) Cramer, J . W., Miller, J . A., and Miller, E. C. (1960) N-
Hydroxylation: A new metabolic reaction observed in the rat with
the carcinogen 2-acetylaminofluorene. J . Biol. Chem. 235, 885-
888.
(13) Calder, I. C., Healey, K., Yong, A. C., Crowe, C. A., Ham, K. N.,
and Tange, J . D. (1978) N-Hydroxyparacetamol - its role in the
metabolism of paracetamol. In Biological Oxidation of Nitrogen
(Gorrod, J . W., Ed.) pp 309-318, Elsevier/North-Holland Bio-
medical Press, New York.
(14) Dahl, G. A., Miller, E. C., and Miller, J . A. (1980) Comparative
carcinogenicities and mutagenicities of vinyl carbamate, ethyl
carbamate, and ethyl N-hydroxycarbamate. Cancer Res. 40,
1194-1203.
(15) Sotomayor, R. E., and Collins, T. F. X. (1990) Mutagenicity,
metabolism, and DNA interactions of urethane. Toxicol. Ind.
Health 6, 71-108.
Replacing one hydrogen of c with methyl reduces the
H-bonding strength and both with methyls eliminates the
five-membered ring and therefore the self-catalysis. The
intramolecular hydrogen bonding leading to self-catalysis
is consistent with the greater stability of 4 and 5 than of
3 and the tendency of 3 (with assistance from the R-effect
factor) to undergo rapid fragmentation, at neutral condi-
tions, via the intermediacy of a metaphosphate analogue.
Although illustrated as intramolecular hydrogen bonding,
the alternative intermolecular hydrogen bonding with
bridging water molecules is more likely in aqueous
medium.
In summary, N-hydroxymethamidophos (3) is a can-
didate metabolite and is toxic to mice as an in vivo AChE
inhibitor yet inactive in vitro as an inhibitor. Methami-
dophos is not a good phosphorylating agent but is
converted to a potent AChE inhibitor (and presumably
phosphorylating agent) in vivo, for which the candidates
are 1-SO and 3 (Scheme 1), the latter probably acting
via the metaphosphate analogue. Formation and rapid
reaction of 3 with AChE in vivo would yield phosphoryl-
ated, aged, and therefore “irreversibly” inhibited AChE,
i.e., MeO(HO)P(O)OAChE, a hypothesis worthy of further
consideration.
(16) Bradford, M. M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72, 248-254.
(17) Ellman, G. L., Courtney, K. D., Andres, V., J r., and Featherstone,
R. M. (1961) New and rapid colorimetric determination of ace-
tylcholinesterase activity. Biochem. Pharmacol. 7, 88-95.
(18) Yoshida, M., Wu, S.-Y., and Casida, J . E. (1994) Reactivity and
stereospecificity of neuropathy target esterase and R-chymo-
trypsin with 2-substituted-4H-1,3,2-benzodioxaphosphorin 2-ox-
ides. Toxicol. Lett. 74, 167-176.
(19) Mahajna, M., Quistad, G. B., and Casida, J . E. (1997) Acephate
insecticide toxicity: Safety conferred by inhibition of the bioac-
tivating carboxyamidase by the metabolite methamidophos.
Chem. Res. Toxicol. 10, 64-69.
(20) Mani, C., Hodgson, E., and Kupfer, D. (1993) Metabolism of the
antimammary cancer antiestrogenic agent tamoxifen. II. Flavin-
containing monooxygenase-mediated N-oxidation. Drug Metab.
Disp. 21, 657-661.
(21) Atta-Asafo-Adjei, E., Lawton, M. P., and Philpot, R. M. (1993)
Cloning, sequencing, distribution, and expression in Escherichia
coli of flavin-containing monooxygenase 1C1. J . Biol. Chem. 268,
9681-9689.
Ack n ow led gm en t. The project described was sup-
ported by Grant R01 ES08762 and Grant P01 ES00049
from the National Institute of Environmental Health
Sciences, NIH. We thank Ernest Hodgson and Nathan
Cherrington (Department of Toxicology, North Carolina
State University, Raleigh, NC) for supplying FMO 1 and
our laboratory colleagues Gary Quistad, Susan Sparks,
and Bachir Latli for assistance and suggestions.
(22) Harger, M. J . P. (1981) O-(Diphenylphosphinyl)hydroxylamine:
preparation and some characteristic chemical reactions. J . Chem.
Soc., Perkin Trans. 1, 3284-3288.
(23) Harger, M. J . P. (1983) Preparation and characterisation of
N-(diphenylphosphinoyl)hydroxylamine, and conversion into O-
sulphonyl derivatives that undergo Lossen-like rearrangement.
J . Chem. Soc., Perkin Trans. 1, 2699-2704.
(24) Harger, M. J . P., and Shimmin, P. A. (1992) N-(Dialkylphos-
phinoyl)hydroxylamines: Preparation using N,O-bis(trimethyl-
silyl)hydroxylamine and migration of simple alkyl groups in the
rearrangements of their O-p-nitrobenzenesulphonates. Tetrahe-
dron Lett. 48, 7539-7550.
Refer en ces
(1) Tomlin, C., Ed. (1994) The Pesticide Manual: A World Compen-
dium, 10th ed., pp 673-674, British Crop Protection Council, The
Bath Press, Bath, U.K.
(2) Quistad, G. B., Fukuto, T. R., and Metcalf, R. L. (1970) Insecti-
cidal, anticholinesterase, and hydrolytic properties of phosphora-
midothiolates. J . Agric. Food Chem. 18, 189-194.
(25) West, R., and Boudjouk, P. (1973) New anionic rearrangements.
XV. 1,2-Anionic rearrangement of organosilylhydroxylamines. J .
Am. Chem. Soc. 95, 3987-3994.
(3) Magee, P. S. (1982) Structure-activity relationships in phospho-
ramidates. In Insecticide Mode of Action (Coats, J . R., Ed.) pp
101-161, Academic Press Inc., New York.
(4) Khasawinah, A. M. A., March, R. B., and Fukuto, T. R. (1978)
Insecticidal properties, antiesterase activities, and metabolism
of methamidophos. Pestic. Biochem. Physiol. 9, 211-221.
(26) Tamura, Y., Ikeda, H., Morita, I., Tsubouchi, H., and Ikeda, M.
(1982) O-Arenesulfonyl-N-alkylhydroxylamines as aminating re-
agents. Chem. Pharm. Bull. 30, 1221-1224.

34 Chem. Res. Toxicol., Vol. 11, No. 1, 1998
Mahajna and Casida
(27) Eto, M. (1974) Organophosphorus Pesticides: Organic and Biologi-
cal Chemistry, pp 162-163, CRC Press, Cleveland, OH.
(28) Grothusen, A., Hardt, J ., Bra¨utigam, L., Lang, D., and Bo¨cker,
R. (1996) A convenient method to discriminate between cyto-
chrome P450 enzymes and flavin-containing monooxygenases in
human liver microsomes. Arch. Toxicol. 71, 64-71.
(29) Forsyth, C. S., and Chambers, J . E. (1989) Activation and
degradation of the phosphorothionate insecticides parathion and
EPN by rat brain. Biochem. Pharmacol. 38, 1597-1603.
(30) Segall, Y., and Casida, J . E. (1982) Oxidative conversion of
phosphorothiolates to phosphinyloxysulfonates probably via phos-
phorothiolate S-oxides. Tetrahedron Lett. 23, 139-142.
(31) J encks, W. P. (1958) The reaction of hydroxylamine with activated
acyl groups. I. Formation of O-acylhydroxylamine. J . Am. Chem.
Soc. 80, 4581-4584.
(33) J encks, W. P. (1958) The reaction of hydroxylamine with activated
acyl groups. II. Mechanism of the reaction. J . Am. Chem. Soc.
80, 4585-4588.
(34) Haake, P., and Allen, G. W. (1980) Biological reactions at
phosphorus. VII. The nature of metaphosphate ion, PO3-, as a
reaction intermediate. Bioorg. Chem. 9, 325-341.
(35) Aubort, J . D., and Hudson, R. F. (1970) The “R-effect” of hydrox-
amic acids. Chem. Commun. 938-939.
(36) Grekov, A. P., and Veselov, V. Ya. (1978) The R-effect in the
chemistry of organic compounds. Russian Chem. Rev. 47, 631-
648.
(37) Sawaki, Y. (1992) Peroxy acids and peroxy esters. In Organic
Peroxides (Ando, W., Ed.) pp 425-478, Wiley, New York.
(38) Richardson, W. H. (1983) Acidity, hydrogen bonding and complex
formation. In The Chemistry of Peroxides (Patai, S., Ed.) pp 134-
135, Wiley, New York.
(32) Marmer, W. N., and Maerker, G. (1972) The preparation and
reactions of novel O-acylhydroxylamines. J . Org. Chem. 37, 3520-
3523.
TX9701135