Identification of a S-hexahydro-1H-azepine-1-carbonyl adduct produced by molinate on rat hemoglobin β2 and β3 chains in vivo

Article abstract of DOI:10.1021/tx015564a

Molinate is a thiocarbamate herbicide used in the rice industry for over 25 years, and regulatory reports have shown that administration of molinate results in reproductive toxicity in male rats. Previous in vitro studies indicate that molinate undergoes oxidative metabolism, forming reactive electrophilic intermediates capable of undergoing nucleophilic addition by protein nucleophiles. On the basis of in vitro studies, carbamylation of an active site serine residue in Hydrolase A has been proposed to be the mechanism responsible for the observed testicular toxicity. The experiments presented here utilize hemoglobin to characterize covalent protein modifications produced in vivo by molinate. Rats were dosed intraperitoneally with molinate as a function of exposure duration. Examination of globin from molinate-treated rats by HPLC demonstrated a new peak in the isolated samples and, when collected and analyzed using MALDI-TOF MS, revealed a 126 Da increase in mass relative to the native β₃ chain. Digestion of the globin using Glu-C and analysis by MALDI-TOF MS revealed two modified peptide fragments at m/z 2743 and 4985 consistent with a 126 Da increase to peptide fragments [122-146] and [102-146] in the unmodified β₂ and β₃ chains of globin. Using selected reaction monitoring LC/MS/MS, S-hexahydro-1H-azepine-1-carbonyl cysteine (HHAC-Cys) was identified in the globin hydrolysates isolated from the molinate-treated rats, but not in the control samples, and the quantity of adduct exhibited a cumulative dose response. These experiments demonstrate the ability of molinate to covalently modify proteins in vivo in a dose dependent manner. For hemoglobin this modification was a carbamylation at Cys-125 similar to the modification produced by disulfiram and N,N-diethyldithiocarbamate. The ability of molinate to covalently modify cysteine residues provides a potential mechanism to account for enzyme inhibition following molinate exposure and suggests that enzymes with cysteine residues in their active site may be inhibited by molinate.

Full text of DOI:10.1021/tx015564a

Chem. Res. Toxicol. 2002, 15, 209-217
209
Id en tifica tion of a S-Hexa h yd r o-1H-a zep in e-1-ca r bon yl
Ad d u ct P r od u ced by Molin a te on Ra t Hem oglobin â2 a n d
â3 Ch a in s in Vivo
Lisa J . Zimmerman,* Holly S. Valentine, Kalyani Amarnath, and
William M. Valentine
Department of Pathology and Center in Molecular Toxicology, Vanderbilt University Medical
Center, Nashville, Tennessee 37232-2561
Received September 7, 2001
Molinate is a thiocarbamate herbicide used in the rice industry for over 25 years, and
regulatory reports have shown that administration of molinate results in reproductive toxicity
in male rats. Previous in vitro studies indicate that molinate undergoes oxidative metabolism,
forming reactive electrophilic intermediates capable of undergoing nucleophilic addition by
protein nucleophiles. On the basis of in vitro studies, carbamylation of an active site serine
residue in Hydrolase A has been proposed to be the mechanism responsible for the observed
testicular toxicity. The experiments presented here utilize hemoglobin to characterize covalent
protein modifications produced in vivo by molinate. Rats were dosed intraperitoneally with
molinate as a function of exposure duration. Examination of globin from molinate-treated rats
by HPLC demonstrated a new peak in the isolated samples and, when collected and analyzed
using MALDI-TOF MS, revealed a 126 Da increase in mass relative to the native â3 chain.
Digestion of the globin using Glu-C and analysis by MALDI-TOF MS revealed two modified
peptide fragments at m/z 2743 and 4985 consistent with a 126 Da increase to peptide fragments
[122-146] and [102-146] in the unmodified â2 and â3 chains of globin. Using selected reaction
monitoring LC/MS/MS, S-hexahydro-1H-azepine-1-carbonyl cysteine (HHAC-Cys) was identified
in the globin hydrolysates isolated from the molinate-treated rats, but not in the control
samples, and the quantity of adduct exhibited a cumulative dose response. These experiments
demonstrate the ability of molinate to covalently modify proteins in vivo in a dose dependent
manner. For hemoglobin this modification was a carbamylation at Cys-125 similar to the
modification produced by disulfiram and N,N-diethyldithiocarbamate. The ability of molinate
to covalently modify cysteine residues provides a potential mechanism to account for enzyme
inhibition following molinate exposure and suggests that enzymes with cysteine residues in
their active site may be inhibited by molinate.
ing administration to rats and that molinate, molinate
sulfoxide, and molinate sulfone demonstrate increasing
potencies for inhibition of liver and testis nonspecific
esterase in vitro (4). Upon the basis of the specificity of
radiolabeled molinate binding and the molecular weight
of the modified protein, it was proposed that molinate
was binding to Hydrolase A through carbamylation of an
active site serine.
In tr od u ction
Molinate (Ordram) is a widely used thiocarbamate
herbicide that is particularly important in the rice
industry (1). Previously, molinate administration was
shown to cause reproductive impairment in male rats
marked by a disruption of spermatogenesis and a reduc-
tion in fertility characterized by a testicular lesion seen
after a single dose (g200 mg/kg) of molinate. Molinate
has been shown to undergo metabolic bioactivation via
oxidation to form the reactive metabolites molinate
sulfoxide and molinate sulfone (2). Both of these electro-
philic species are capable of forming glutathione conju-
gates and can be further metabolized into the mercap-
turates and excreted in the urine (2). Previous studies
have supported the sulfoxidation of molinate as the
bioactivation pathway responsible for the testicular toxic-
ity observed following molinate administration (3). In-
hibition of Hydrolase A, a carboxylesterase required for
the production of testosterone in the rat, is the proposed
mechanism of testicular toxicity. Evidence supporting
this mechanism was derived from observations that
molinate inhibits esterase activity in Leydig cells follow-
Disulfiram, a dithiocarbamate, carbamylates proteins
in vivo, presumably through generation of an S-methyl
N,N-diethylthiocarbamate sulfoxide or sulfone metabolite
(5). Consequently, administration of disulfiram to rats
results in high levels of carbamylation on hemoglobin
that is selective for Cys-125 located on the â2 and â3
globin chains. Similarly, in vivo carbamylation of the
active site cysteine residue of mitochondrial aldehyde
dehydrogenase (ALDH)1 by disulfiram has been observed
and proposed to be responsible for inhibition of this
enzyme (6). Because disulfiram and molinate are both
metabolized to thiocarbamate sulfoxides and sulfones,
molinate may produce similar protein modifications to
those observed for disulfiram. This suggests that hemo-
globin may be a good surrogate for characterizing protein
modifications produced by molinate in vivo as well.
* To whom correspondence should be addressed. E-mail:
lisa.j.zimmerman@vanderbilt.edu.
10.1021/tx015564a CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/23/2002

210 Chem. Res. Toxicol., Vol. 15, No. 2, 2002
Zimmerman et al.
Ch em ica l Syn th esis. Molin a te (S-Eth yl-h exa h yd r o-1H-
a zep in e-1-ca r both ioa te) (1). Hexamethyleneimine (1.13 mL,
10 mmol), ethanol (10 mL), and 10 N NaOH (1 mL) were taken
up in a 25 mL two-neck flask and fitted with a Dewar condenser
containing dry ice and 2-propanol. The flask was stirred and
cooled in a dry ice/2-propanol bath (-70 °C), and carbonyl sulfide
(COS) (1 mL) was condensed into the flask. The second neck
was sealed, and the temperature of the bath increased to 0-10
°C and maintained in that temperature range while keeping
the Dewar condenser cold while COS evaporated and condensed
into the reaction. After 4 h, the condenser was removed, and
iodoethane (0.88 mL, 11 mmol) was added with stirring for 2 h,
after which the EtOH was removed, and the residue distributed
between the water (10 mL) and chloroform (20 mL) layers. The
organic layer was separated and dried, and the solvent was
The investigation presented here was performed to
extend previous in vitro studies through determining if
molinate exposure results in covalent protein modifica-
tions in vivo and to identify the adducts generated and
the amino acid residues modified. To accomplish this, rats
were exposed to molinate as a function of dose adminis-
tered in 4 and 11 day durations and hemoglobin isolated
and analyzed using mass spectrometry and HPLC.
Characterization of the protein modifications produced
by molinate will aid in evaluating potential biological
effects of molinate in addition to testicular toxicity, and
may provide a basis for the development of biomarkers
of exposure or effect.
1
removed using a rotary evaporator. H NMR (CDCl3) δ 1.27 (t,
Ma ter ia ls a n d Meth od s
3H, CH3), 1.51 (m, 4H, CH2), 1.70 (m, 4H, CH2), 2.88 (q, 2H,
CH2CH3), 3.41 (t, 2H, ring CH2-N), 3.52 (t, 2H, ring CH2-N).
13C NMR (CDCl3) δ 15.8 (CH3), 24.9 (CH2CH3), 27.4 and 27.6
(CH2), 28.3 and 28.8 (CH2), 47.6 and 48.0 (CH2N), 168.2 (Cd
O).
Ch em ica ls. Sigma Chemical Co. (St. Louis, MO) was the
source for corn oil and Endoproteinase Glu-C (from Staphylo-
coccus aureus strain V8). J . T. Baker (Phillipsburg, NJ ) was the
source for poly(ethylene glycol) 200 (PEG). The Alzet osmotic
minipumps (2 mL capacity, 4 week delivery) were obtained from
Durect Corporation (Cupertino, CA). Unless otherwise specified,
additional chemicals were also obtained from commercial sources.
An im a ls. This study was performed in accordance with the
National Institutes of Health’s Guide for Care and Use of
Laboratory Animals and was approved by the Institutional
Animal Care and Use Committee. For animal exposures, male
Sprague Dawley rats, 16 weeks old and 370-450 g, (Harlan,
Sprague Dawley, Indianapolis, IN) were used. Rats were housed
in a room on a 12 h diurnal light cycle and given rodent chow
and water ad libitum.
Molin a te Exp osu r es. Prior to molinate administration, male
rats (n ) 4) were anesthetized and 0.5 mL of whole blood was
collected in a heparinized syringe from the tail vein and globin
isolation performed. A once daily dose of 100 mg/kg molinate
in corn oil (320 µL of molinate added to 1.78 mL corn oil) was
then given by intraperitoneal injection for a period of 4 days.
On day 5, the rats were anesthetized and blood was drawn again
for globin isolation.
S-Eth yl-h exa h yd r o-1H-a zep in e-1-ca r both ioa te su lfon e
(2). A solution of molinate (0.94 g, 5 mmol) in dichloromethane
(10 mL) was cooled in an ice bath, and m-chloroperoxybenzoic
acid (2.70 g, 11 mmol) in the same solvent (25 mL) was added
with stirring. The reaction was stirred at 0 °C for 4 h, the solid
was filtered and the filtrate was washed with saturated
NaHCO3 (2 × 25 mL), and then dried and concentrated using a
rotary evaporator. The product was purified using column
chromatography (hexane; 9:1 hexane/ethyl acetate; 5:1 hexane/
ethyl acetate) and fractions containing the product giving a GC
peak at 10.5 min were combined and the solvent removed under
reduced pressure yielding a colorless liquid. 1H NMR (CDCl3) δ
1.36 (t, 3H, CH3), 1.57 (m, 4H, ring CH2), 1.78 (m, 4H, ring CH2),
3.30 (q, 2H, CH2CH3), 3.49 (t, 2H, CH2-N), 3.84 (t, 2H, CH2-
N). 13C NMR (CDCl3) δ 6.7 (CH3), 26.0, 26.3, 27.3, 28.8 (ring
CH2), 46.1 (CH2CH3), 46.8 and 48.8 (CH2-N), 160.6 (CdO).
S-(Hexa h yd r o-1H-a zep in e-1-ca r bon yl)cystein e (HHAC-
Cys) (3) was prepared according to previously published
procedures with slight modifications (5, 7). Briefly, the pH of a
solution of L-cysteine (0.52 g, 4 mmol) in water was adjusted to
8 with 1 N NaOH and stirred with molinate sulfone (0.44 g, 2
mmol) in methanol for 24 h. After removing the ACN, purifica-
tion was achieved by HPLC by injecting the reaction mixture
onto a C18 PRP-1 column (Hamilton, 70 µm, 7.0 × 305 mm,
100 Å). The desired product was eluted at a flow rate of 2.0 mL/
min using a linear gradient of 10-90% B over 10 min and then
maintained at 90% B for 9 min before returning to initial
conditions. Solvent A contained 5 mM formic acid in water, and
solvent B contained 5 mM formic acid in acetonitrile. The
separation was monitored at 214 nm, and fractions containing
the sulfone were combined and lyophilized leaving a white solid.
1H NMR (CDCl3) δ 1.48 (m, 4H, CH2), 1.65 (m, 4H, CH2), 3.26
(q, 2H, CH2), 3.64 (m, 4H, CH2), 3.96 (q, 1H, CH). 13C NMR
(CDCl3) δ 26.4, 26.8, 27.36 and 27.07 (ring CH2), 30.7 (S-CH2),
48.4 and 48.7 (ring CH2), 55.2 (CH-NH2), 169.2 (CdO), 172.6
(N-CdO). ESI-MS m/z 247 (M + H)+.
S-Meth yl-1-p ip er id in eca r both ioa te (4). To piperidine (20
mmol, 1.98 mL) in ethanol (20 mL) was added 10 N NaOH (2
mL) in a 25 mL two-neck flask and fitted with a Dewar
condenser containing dry ice and 2-propanol. The flask was
stirred and cooled in a dry ice/2-propanol bath (-70 °C) and
COS (2 mL) was condensed into the flask. The second neck was
sealed, and the temperature of the bath raised to 10 °C and
maintained at that temperature while keeping the Dewar
condenser cold while COS evaporated and condensed into the
reaction. After 4 h, the condenser was removed and iodomethane
(1.4 mL, 22 mmol) was added with stirring for 2 h after which
the EtOH was removed, and the residue distributed between
water (10 mL) and chloroform (20 mL) layers. The organic layer
was separated, dried and evaporated. 1H NMR (CDCl3) δ 1.53
(m, 6H), 2.25 (s, 3H, CH3), 3.41 (m, 4H, CH2-N). 13C NMR
For the second exposure experiment, four week Alzet osmotic
minipumps delivering 2.5 µL/h of molinate in PEG 200 (389 mg/
mL) for an effective dose of 1 mmol/kg/day (187 mg/kg/day) were
surgically implanted in the abdomen of three male rats. Whole
blood (0.5 mL) was collected and globin isolation performed on
day 11 following implantation.
Globin Isola tion . A total of 0.5 mL of whole blood was
collected in a heparinized syringe at each bleeding time and then
centrifuged at 3000g for 5 min to separate the plasma from the
red cells. A 5 mM phosphate-buffered saline solution (pH 7.4)
containing 150 mM NaCl was added in equal volume to the red
cells, resuspended, and centrifuged at 3000g. The supernatant
and buffy coat were discarded and the washing procedure
repeated twice more. The washed red cells were then lysed with
a 2× volume of 5 mM phosphate buffer (pH 7.4) and centrifuged
at 20000g for 25 min. The hemolysate (supernatant) was
removed and mixed with 100 mL of 1 M ascorbic acid and added
dropwise to 10 mL of cold 2.5% oxalic acid in acetone and
allowed to precipitate for 15 min. The mixture was then
centrifuged at 12000g for 10 min, the supernatant aspirated,
and the globin washed by adding 5 mL of acetone, resuspending
with a metal spatula, and centrifuging at 12000g for 10 min.
The supernatant was aspirated and the globin pellet dried under
a stream of nitrogen and then stored at -78 °C.
1 Abbreviations: ACN, acetonitrile; ALDH, aldehyde dehydrogenase;
m-CPBA, meta-chloroperoxybenzoic acid; CID, collision-induced dis-
sociation; COS, carbonyl sulfide; DSF, disulfiram; ESI, electrospray
ionization; HHAC-Cys, S-hexahydro-1H-azepine-1-carbonyl cysteine;
MeOH, methanol; MALDI-TOF MS, matrix-assisted laser desorption
ionization time-of-flight mass spectrometry; PIP-Cys, S-piperidine-1-
ylcarbonyl cysteine; SA, sinapinic acid; SRM, selected reaction moni-
toring; TFA, trifluoroacetic acid.

Hemoglobin Adduct in Rats Exposed to Molinate
Chem. Res. Toxicol., Vol. 15, No. 2, 2002 211
(CDCl3) δ 13.2 (CH3), 24.8 (NCH2CH2), 26.0 (CH2), 46.6 (NCH2),
167.6 (CdO).
a Costar Spin-X centrifuge spin filter (0.2 µm pore size). Samples
were stored at -20 °C until analysis.
S-Meth yl-1-p ip er id in eca r both ioa te su lfon e (5). A solu-
tion of S-methyl-1-piperidinecarbothioate (0.71 g, 4.4 mmol) in
dichloromethane (10 mL) was cooled in an ice bath and m-CPBA
(2.4 g) in the same solvent (20 mL) was stirred overnight at
room temperature. The solid was filtered off and the filtrate was
washed with NaHCO3. The organic layer was dried and evapo-
rated and purified by column chromatography (hexane; 9:1
hexane/ethyl acetate; 5:1 hexane/ethyl acetate; 1:1 hexane/ethyl
acetate). Fractions containing the product giving a GC peak at
9.2 min were combined and evaporated. 1H NMR (CDCl3) δ 1.67
(m, 6H, ring CH2), 3.1 (s, 3H, methyl), 3.33 (m, 2H, CH2-N),
3.65 (m, 2H, CH2-N). 13C NMR (CDCl3) δ 23.9, 25.3, 25.9, 39.6
(CH3), 45.7 and 46.2 (CH2-N), 159.6 (CdO).
Samples for LC/MS/MS analysis were prepared in a similar
manner using 10 mg of protein and reconstituting the hydro-
lyzed sample in 100 µL ACN/H2O/formic acid (10:90:0.02%).
Ma ss Sp ectr om etr y. MALDI-TOF MS analysis of intact
proteins and digests were performed on a Voyager Elite (Per-
Septive Biosystems, Inc.) time-of-flight mass spectrometer. The
ion acceleration was 25 kV, the grid voltage 95% for intact
proteins and 93% for peptides, and the guide wire voltage 0.1%.
Mass spectra were acquired as the sum of ions generated by
irradiation of the sample with 100-300 laser pulses and
processed using GRAMS/386 software (Galactica Industries
Corp., Salem, NH). A saturated solution of sinapinic acid (SA)
in ACN/H2O/TFA (50:49:0.1 v/v/v) was used as the matrix. For
analysis of intact proteins, 2 µL of a globin mixture (1 mg/mL)
was added to 2 µL of the SA matrix and applied to the plate by
the dried-droplet technique and air-dried. Myoglobin was used
as an internal standard with an m/z of 16952.5. For analysis of
the peptide mixtures following enzymatic digestion of intact or
purified protein fractions, 2-4 µL of the mixture was added to
an equivalent volume of the SA matrix and applied to the
sample plate in a similar manner. The mass of two known
peptide fragments derived from the â3 chain, [27-43] and [27-
90] with m/z of 2098.5 and 7047.1, respectively, were used as
internal calibration standards (5).
S-(P ip er id in e-1-ylca r bon yl)cystein e (P IP -Cys) (6). The
pH of a solution of L-cysteine (0.52 g, 4 mmol) in water was
adjusted to 8 with 1 N NaOH and stirred with S-methyl-1-
piperidinecarbothioate sulfone (0.38 g, 2 mmol) in MeOH (20
mL) for 24 h. After removing the MeOH, purification was
accomplished by HPLC by injecting the reaction mixture onto
a PRP-1 column (Hamilton, 70 µm, 7.0 × 305 mm, 100 Å). The
desired product was eluted at a flow rate of 2.0 mL/min using
a linear gradient of 10 to 90% B over 20 min using the same
solvent system for (3). The eluant was monitored at 214 nm and
fractions containing the product were lyophilized, leaving a
1
white solid. H NMR (D2O) δ 1.54 (m, 6H, ring CH2), 3.29 (dd,
2H, S-CH2), 3.50 (m, 4H, ring CH2), 3.93 (q, 1H, CH-NH2). 13C
NMR (D2O) δ 24.0, 25.5, 31.0, 45.8, 47. 9, 55.3, 167.8 (NCdO),
172.9 (CdO).
Tandem mass spectrometry was performed on a Waters 2690
HPLC system coupled with
a Finnigan TSQ 7000 triple-
quadrupole ESI/MS (Finnigan, San J ose, CA) equipped with a
nano-ESI ion source. Ten microliters of the globin acid hydroly-
sate was loaded into the nanospray capillary (Picotips, New
Objectives, Cambridge, MA). The heated capillary was operated
at 20 V and 100 °C, tube lens voltage set to 80 V and the needle
voltage 1000 V. Collision-induced dissociation (CID) occurred
in Q2 with argon (2.5 mT) as the carrier gas with collision
energy varied between 5 and 30 eV.
Ch r om a togr a p h y. Analysis of intact globin samples was
performed on a Waters 2690 HPLC system. Globin (1.8-2.5 mg)
from nonexposed and molinate-exposed rats was dissolved in
0.1% TFA (750 µL). The absorbance was measured at 280 nm
and diluted as necessary to produce a final absorbance of 1. Fifty
microliters was injected onto a C4 RP column (4.6 mm × 150
mm, 5 µm, 300 Å, Phenomenex, Torrance, CA). Globin samples
were eluted at a flow rate of 1 mL/min with a linear gradient
from initial conditions of 56% A [solvent A ) ACN/H2O/TFA
(20:80:0.1%)] and 44% B [solvent B ) ACN/H2O/TFA (60:40:
0.08%)] to 30% A and 70% B over 30 min while monitoring at
214 nm.
Liquid chromatography tandem mass spectrometry (LC/MS/
MS) analysis of HHAC-Cys with selected reaction monitoring
(SRM) was performed using a Waters 2690 HPLC system
coupled with a Finnigan TSQ 7000 triple-quadrupole ESI/MS
(Finnigan, San J ose, CA). A PE Biosystems (Foster City, CA)
785A Programmable Absorbance UV detector was used and set
to monitor at 214 nm. The electrospray voltage was maintained
at 4 kV and the capillary temperature held at 200 °C. Collision-
induced dissociation (CID) occurred in Q2 at a collision energy
of 20 eV with the collision gas (argon) at 2.5 mT. Solvent A
contained 5 mM formic acid in water and solvent B contained 5
mM formic acid in acetronitrile. LC separations were performed
on a C18 PRP-1 column (3 µm, 2.1 × 100 mm, 100 Å, Hamilton,
Reno, NV). Separations were achieved using initial conditions
of 10% B (90% A) held for 10 min followed by a linear gradient
to 90% B (10% A) over 5 min and held for 6 min before returning
to initial conditions. Equilibration with 10% B (90% A) was
allowed for 13 min and total run time was 35 min. Twenty-five
microliters of the sample treated with 5 µL of the internal
standard (80 ng/µL) was injected and the eluant from 2-9 min
directed into the mass spectrometer. SRM experiments used to
detect HHAC-Cys monitored the m/z 247f126 transition and
the m/z 233f112 transition monitored for PIP-Cys, the internal
standard. The scan width for the daughter ions was 1 amu and
total cycle time was 2 s. The ratios of the resulting peak areas
were used to calculate the amount of HHAC-Cys in the acid
hydrolysates by reference to a calibration curve. A calibration
curve was generated by dissolving varying amounts of HHAC-
Cys in solvent A and adding a fixed amount of PIP-Cys so each
25 µL injection would contain 0-1000 ng of HHAC-Cys and 100
ng of PIP-Cys. The ratio of the area of the analyte to the
internal standard was plotted against the concentration of
Globin R- and â-chains were separated using HPLC by
dissolving globin collected from nonexposed and molinate-
exposed rats in solvent A to a final concentration of 10 mg/mL
and loading 50-100 µL onto a C4 Waters Delta Pak column
(15 µm, 3.9 × 300 mm, 300 Å). Protein fractions were eluted at
a flow rate of 2 mL/min with a linear gradient from initial
conditions of 56% A [solvent A ) ACN/H2O/TFA (20:80:0.1%)]
and 44% B [solvent B ) ACN/H2O/TFA (60:40:0.08%] to 40% A
and 60% B over 30 min while monitoring at 214 nm. Protein
fractions were collected, and ACN was removed and lyophilized
before analysis with MALDI-TOF MS.
Separations and collection of fractions were performed on a
Shimadzu LC-10ADvp connected to a SPD-10A detector. The
software used in the analysis of the data was HP Chemstation.
En zym a tic Digestion . All enzymatic digestions on globin
preparations and purified â-chain fractions were performed in
solution by adding 5 µL of globin (1 mg/mL) or purified â-chains
(0.3-0.5 mg/mL) to 10 µL 100 mM NH4HCO3 buffer (pH 6.5)
and 1.5 µg Glu-C. Samples were digested for 2 h at 37 °C.
Following digestion, 2-4 µL of the digest was added to an
equivalent amount of a saturated sinapinic acid (SA) matrix
solution and applied to a MALDI probe using the dried droplet
method.
Acid Hyd r olysis. For nano-ESI MS experiments, approxi-
mately 1 mg of globin was added to a hydrolysis tube, placed in
a hydrolysis chamber containing 200 µL of 6 N HCl, flushed
with argon gas, and hydrolyzed for 18 h at 110 °C. Following
completion of hydrolysis, samples were reconstituted in 100 µL
of MeOH/H2O/formic acid mixture (50:49:1%) and filtered using
HHAC-Cys in the sample to generate a standard curve (r2
0.992, y ) 0.005142x + 0.228125).
)

212 Chem. Res. Toxicol., Vol. 15, No. 2, 2002
Zimmerman et al.
F igu r e 1. HPLC analysis of rat globin isolated by precipation
followed by separation using a C4 reversed-phase column and
detection by UV absorption at 214 nm. (A) Representative
chromatogram of nonexposed globin showing a combined R-peak
eluting between 18 and 20 min and two peaks resulting from
the unresolved four native â-chains eluting between 22 and 25
min. (B) Globin preparations isolated from molinate exposed
animals produced an additional late eluting peak (26-27 min).
Resu lts
In ta ct Globin An a lyis by MALDI-TOF MS. The
HPLC chromatograms obtained from the globins of
molinate-exposed and nonexposed rats are shown in
Figure 1. HPLC analysis of both the nontreated and
molinate-treated globin samples produced a single peak
eluting between 18 and 20 min corresponding to the two
native R-chains, and two peaks eluting between 22 and
25 min corresponding to the four native â-chains. The
globin isolated from molinate-exposed rats contained an
additional late eluting peak between 26 and 27 min that
was not seen in the controls. MALDI-TOF MS analysis
of the control globin samples did not resolve individual
peaks for each globin chain but appeared to resolve the
two R-chains at m/z 15 150 and 15 197 with a single peak
for the â-chains at m/z 15 849 (Figure 2A). Analysis of
the globin isolated from molinate-exposed rats also
produced peaks at m/z 15 150, 15 197, and 15 849 with
an additional signal at m/z 15 975 not seen in the control
(Figure 2B). Isolation of the late eluting HPLC peak
between 26 and 27 min present in molinate-exposed
samples followed by MALDI-TOF MS demonstrated it to
have an m/z of 15 975 (Figure 2C). MALDI-TOF MS
analysis of the fraction eluting between 22 and 25 min
produced a signal at m/z 15 990 in samples from exposed
rats (Figure 2D) but not from control samples (data not
shown). Myoglobin was used as an internal standard (m/z
16 952.5) and SA matrix adducts 200 or 225 Da larger
in mass than the protein signals were also observed.
Glu -C Digests of Globin . Globin samples from ex-
posed and nonexposed rats were digested in solution with
F igu r e 2. MALDI-TOF MS spectra of (A) control globin
showing broad signals at m/z 15 150, 15 197, and 15 849
corresponding to the two native R- and four native â-chains,
respectively. (B) Spectra of globin isolated from molinate-treated
rats contained a signal at m/z 15 975 additional to those
observed in the nontreated globins. (C) A spectrum produced
by the late eluting HPLC fraction (26-27 min) isolated from
exposed globin preparations demonstrating an m/z of 15 975
consistent with the addition of a 126 Da S-hexahydro-1H-
azepine-1-carbonyl adduct to the â3 chain. (D) In the fraction
eluting between 22 and 25 min isolated from molinate exposed
globins a signal at m/z 15 990 consistent with the addition of a
126 Da S-hexahydro-1H-azepine-1-carbonyl adduct to the â2
chain could be detected. Myoglobin was used as an internal
standard (m/z 16 952.2) and SA matrix adducts can been seen
as signals 200 or 225 Da larger than the parent protein signals
described above.
Glu-C for 2 h and analyzed using MALDI-TOF MS. The
peptide fragments originating from the â3 chain of globin
could then be assigned (Figure 3). Two signals from the
peptide fragments with an m/z at 2743 and 4985, corre-

Hemoglobin Adduct in Rats Exposed to Molinate
Chem. Res. Toxicol., Vol. 15, No. 2, 2002 213
F igu r e 3. MALDI-TOF MS spectra of Glu-C digests in the regions of m/z 2000-3000 and 4500-5100. (A) Spectra of globin isolated
from a nonexposed rat showing the predicted peptide fragments. (B) Spectra of globin isolated from molinate-treated rats produced
two additional peptide fragments at m/z 2743 and 4985, corresponding to the addition of 126 Da to the peptides [122-146] and
[102-146], respectively. (C) Spectra of the isolated HPLC fraction eluting between 26 and 27 min containing the modified â3 chain.
Both modified fragments observed in panel B are present although the signal at m/z 2617 corresponding to the native peptide [122-
146] is not present. For all spectra, peptides at [27-43] (m/z 2098.5) and [27-90] (m/z 7047.1 not shown) were used for internal
calibration.
sponding to the addition of 126 Da to the peptides [122-
146] and [102-146], respectively, were observed in the
digest of the molinate-exposed rats (Figure 3B) but were
not present in the nontreated globin (Figure 3A). Frac-
tions obtained from the peak eluting between 26 and 27
min were digested and while no signal at m/z 2617
corresponding to the native [122-146] peptide was
observed, the two signals at m/z 2743 and 4985 appeared,
corresponding to the modified fragments (Figure 3C).
Na n o-ESI-MS of Acid Hyd r olyzed Globin s. A CID
of the parent ion of the synthetic authentic HHAC-Cys
at m/z 247 produced fragments ions at m/z 230, 201, 160,
126, 98, 83, and 55 (Figure 4A). The ion at m/z 126 can
be attributed to the loss of the carbonyl hexamethylene-
imine moiety from the parent compound. Although the
control hydrolysate produced several fragments from the
parent of m/z 247 using identical CID conditions, there
were no ions detected at m/z 247, 230, 201, 126, 98, 83,
or 55, characteristic of the HHAC-Cys adduct (Figure 4B).
Acid hydrolysates of globin from molinate-treated rats
contained the same fragment ions observed for the
synthetic HHAC-Cys standard in addition to those
observed for the control samples (Figure 4C).
An a lysis of Globin s u sin g LC/MS/MS. Acid hydroly-
sates of nonexposed and molinate-exposed globin were
analyzed for the HHAC-Cys adduct using selected reac-
tion monitoring (SRM). The retention times of synthetic
HHAC-Cys and the internal standard, PIP-Cys, were 6.5
and 4 min, respectively, when chromatographed on a C18
PRP-1 column (Figure 5A). Control hydrolysates did not
contain a peak for the HHAC-Cys adduct at 6.5 min,
although the internal standard could be seen (Figure 5B).
Molinate-treated globin hydrolysates did, however, con-
tain a peak having the same retention time as the

214 Chem. Res. Toxicol., Vol. 15, No. 2, 2002
Zimmerman et al.
F igu r e 4. Product ions obtained from CID of the [MH+] ion
m/z 247 of S-hexahydro-1H-azepine-1-carbonyl cysteine. (A) A
spectrum of the synthetic standard S-hexahydro-1H-azepine-
1-carbonyl cysteine (HHAC-Cys) showing characteristic frag-
ments at m/z 230, 201, 160, 126, 98, 83, and 55. (B) A spectrum
obtained from control hydrolysate showing one major fragment
at m/z 116. (C) A spectrum obtained from molinate-exposed
hydrolysates showing a combination of the fragments present
in the synthetic standard and control hydrolysate. The ion at
m/z 126 is attributed to loss of the S-hexahydro-1H-azepine-1-
carbonyl adduct and is observed in the standard and exposed
rats (A and C) but not in the control (B).
F igu r e 5. LC/MS/MS with SRM of globin hydrolysates. The
internal standard, S-(piperidine-1-ylcarbonyl) cysteine (PIP-Cys
m/z 233), is shown in the lower trace of each chromatogram with
a retention time of 4 min. The presence of HHAC-Cys, with a
retention time of 6.5 min, is shown in the upper trace of each
chromatogram for the synthetic standard (A) and the molinate-
exposed globin (C) while HHAC-Cys is not detected in the
control (B).
synthetic standard (Figure 5C). The quantity of adduct
detected in the samples is shown as a function of the
cumulative exposure (Figure 6). Estimates of the amounts
of adduct were calculated from the calibration curve
generated from known quantities of the authentic stan-
dard. Analysis performed on 10 mg samples demon-
strated that rats dosed with molinate for 4 days at 100
mg/kg/day (400 mg/kg total dose) corresponded to 720 (
99 ng/10 mg of protein of the adduct while 1233 ( 41
ng/10 mg of protein adduct was detected in the rats dosed
for 11 days at 187 mg/kg/day (2000 mg/kg total dose).

Hemoglobin Adduct in Rats Exposed to Molinate
Chem. Res. Toxicol., Vol. 15, No. 2, 2002 215
alleviating the need for HPLC separation of the two
subgroups of chains prior to digestion (8, 9). Using Glu-C
enzymatic digestion followed by MALDI-TOF MS analy-
sis of the digests, the 126 Da increase was localized to
two peptides of â2 and â3, [122-146] and the partial
digestion product [102-146], both containing a reactive
cysteine at residue 125. To confirm the structure of the
proposed covalent modification at Cys-125, hydrolysates
from the molinate-treated globins were examined by ESI-
MS and found to produce fragments identical to those
produced from the HHAC-Cys synthetic standard, whereas
no analogous fragments were observed in the control
hydrolysates. LC/MS/MS analysis using SRM to detect
the HHAC-Cys adduct following hydrolysis indicated that
the quantity of the HHAC-Cys adduct displayed a
cumulative dose-response over the duration of molinate
exposure with adducts detected as early as 4 days. These
results are consistent with molinate being bioactivated
to a sulfoxide or sulfone intermediate that undergoes
nucleophilic addition by Cys-125 of globin to yield an acid
stable carbamoyl adduct (Scheme 2). Additional metabolic
pathways have been reported for molinate that are not
expected to result in protein modifications. Oxidative
metabolism of molinate to produce ring-hydroxylated
metabolites that are typically excreted unchanged or as
glucuronide conjugates are generated, and hydrolysis of
the sulfone to liberate hexahydroazepine may also occur
(2, 10). The molinate sulfoxide and sulfone can also form
glutathione conjugates, which, after further metabolism
are excreted in the urine as mercapturic acids (2).
Previous in vitro studies using microsomal prepara-
tions have provided evidence that covalent binding of
radiolabeled molinate and its metabolites occurs prefer-
entially on a single carboxylesterase in the liver and
testis (3). The relative potencies for carboxylesterase
inhibition observed for the sulfoxide and sulfone deriva-
tives of molinate were consistent with the inhibition
occurring through the generation of an S-hexahydro-1H-
azepine-1-carbonyl adduct (1). The further observation
that prior incubation with phenylmethanesulfonyl fluo-
ride (PMSF) prevented molinate binding was interpreted
to support the covalent modification occurring at the
active site (3). In contrast, the data derived from the
present in vivo study examining globin digests demon-
strates the formation of the analogous adduct on Cys-
125 in the absence of serine modification. Most likely this
discrepancy arises due to the high reactivity of these two
specific residues on their respective proteins (11, 12).
F igu r e 6. HHAC-Cys formation as a function of cumulative
dose (mg/kg/day × days of exposure) for globin samples obtained
from control (O) and molinate-treated (9) rats. Dose-response
curves relate to the quantity of HHAC-Cys detected for control,
4 day (400 mg/kg total dose), and 11 day (2000 mg/kg total dose)
exposures. Normalized values for HHAC-Cys responses were
obtained by dividing the mass spectrometric response for
HHAC-Cys by the internal standard PIP-Cys and the amount
of protein hydrolyzed multiplied by 100 to give a dimensionless
number. Error bars represent standard errors (n ) 3).
Sch em e 1. Syn th esis of HHAC-Cys a n d P IP -Cys
Discu ssion
HPLC chromatograms of globin samples isolated from
molinate-exposed rats contained a late eluting peak
additional to the native R- and â-chains. Isolation and
MALDI-TOF MS analysis of the late eluting peak re-
vealed a mass corresponding to the â3 chain plus 126 Da,
consistent with the formation of a S-hexahydro-1H-
azepine-1-carbonyl adduct. MALDI-TOF MS also re-
vealed a small signal at m/z 15 990, indicative of a
similarly modified â2 chain. Glu-C has been shown to
digest â-chains with minimal digestion of the R-chains,
Sch em e 2. P r op osed Meta bolic P a th w a y for S-Hexa h yd r o-1H-a zep in e-1-ca r bon yl Ad d u ct F or m a tion

216 Chem. Res. Toxicol., Vol. 15, No. 2, 2002
Zimmerman et al.
Although, there is some uncertainty regarding the exact
site of the protein modification in the microsomal prepa-
ration in that the structure and location of the modifica-
tion was not determined and PMSF has been reported
to also modify cysteine residues (13). Regardless, the
current data is consistent with the ability of molinate to
potentially alter protein function through the modifica-
tion of cysteine residues in addition to the serine based
modification proposed previously.
to thiocarbamate herbicides. However, since human
hemoglobin lacks the highly reactive Cys-125 exclusive
to rat hemoglobin, further experiments will be required
to determine the reactivity of molinate and its metabo-
lites toward human hemoglobin.
Ack n ow led gm en t. L.J .Z. would like to thank V.
Amarnath for the guidance in the synthesis of the
compounds used for this study. We also would like to
thank the Vanderbilt University Mass Spectrometry
Resource for the use of the mass spectrometer and Lisa
Manier for her technical assistance. These studies were
supported by NIEHS Grant ES06387 and Center in
Molecular Toxicology Grant P30 ES00267. In addition,
L.J .Z. was supported, in part, by NRSA Grant F32
AA05569-02.
The metabolism of DSF has been characterized and,
similar to molinate, is thought to be bioactivated through
a series of intermediates to form S-methyl N,N-dieth-
ylthiocarbamate sulfoxide (DETC-MeSO), which is thought
to be the active metabolite of DSF and ultimate inhibitor
of ALDH (14-17). Recently DSF has been shown to
covalently modify ALDH in vivo at the cysteine active
site through carbamylation, increasing the mass of the
enzyme by 100 Da (6). Disulfiram has also been shown
to covalently modify rat hemoglobin in vivo, generating
a stable N,N-diethylthiocarbamate adduct through car-
bamylation of Cys-125 on the â2 and â3 chains (5). It is
interesting to note that, in addition to generating analo-
gous covalent modifications on globin, molinate has also
been shown to inhibit ALDH, to increase levels of
acetaldehyde, and to produce a disulfiram-like ethanol
reaction in rats challenged with ethanol (18). The results
presented here for globin suggest that inhibition of ALDH
by molinate may also occur through carbamylation of an
active site cysteine by the sulfoxide metabolite. The
ability of molinate to undergo bioactivation to form
reactive metabolites which can carbamylate proteins in
vivo, similar to DSF, suggests that other biological effects
associated with DSF may also be relevant to molinate.
In particular, administration of DSF has been shown to
be hepatotoxic and produce a peripheral neuropathy that
targets Schwann cells, resulting in demyelination (19).
Although the literature regarding the neurotoxicity as-
sociated with molinate is limited, there have been reports
revealing an increased incidence of nerve degeneration,
Schwann cell hyperplasia, and increased frequency of
eosinophilic bodies in the spinal cord and medulla (20).
If protein carbamylation is a contributing mechanism in
the neurotoxicity of disulfiram then molinate may also
be neurotoxic through a similar mechanism and may
display a greater potency given the fact that fewer
bioactivation steps are required to generate the molinate
sulfoxide or sulfone.
Refer en ces
(1) J ewell, W. T., Hess, R. A., and Miller, M. G. (1998) Testicular
toxicity of molinate in the rat: Metabolic activation via sulfoxi-
dation. Toxicol. Appl. Pharmacol. 149, 159-166.
(2) DeBaun, J . R., Bova, D. L., Tseng, C. K., and Menn, J . J . (1978)
Metabolism of 14C-Ordram(molinate) in the rat. 2. Urinary
metabolite identification. J . Agric. Food Chem. 26, 1096-1104.
(3) J ewell, W. T., and Miller, M. G. (1998) Identification of a
carboxylesterase as the major protein bound by molinate. Toxicol.
Appl. Pharmacol. 149, 226-234.
(4) Ellis, M. K., Richardson, A. G., Foster, J . R., Smith, F. M.,
Widdowson, P. S., Farnworth, M. J ., Moore, R. B., Pitts, M. R.,
and Wickramaratne, G. A. de S. (1998) The reproductive toxicity
of molinate and metabolites to the male rat: Effects of testoster-
one and sperm morphology. Toxicol. Appl. Pharmacol. 151, 22-
32.
(5) Erve, J . C., J ensen, O. N., Valentine, H. S., Amarnath, V., and
Valentine, W. M. (2000) Disulfiram generates a stable N,N-
diethylcarbamoyl adduct on Cys-125 of rat hemoglobin. Chem.
Res. Toxicol. 13, 237-244.
(6) Shen, M. L., J ohnson, K. L., Mays, D. C., Lipsky, J . J ., and Naylor,
S. (2001) Determination of in vivo adducts of disulfiram with
mitochondrial aldehyde dehydrogenase. Biochem. Pharmacol. 61,
537-545.
(7) Tjeerdema, R. S., and Crosby, D. G. (1988) Disposition, biotrans-
formation, and detoxification of molinate (Ordram) in whole blood
of the common carp (Cyprinus carpio). Pestic. Biochem. Physiol.
31, 24-35.
(8) Houmard, J ., and Drapeau, G. R. (1972) Stapyloccoccal Pro-
tease: A proteolytic enzyme specific for glutamoyl bonds. Proc.
Natl. Acad. Sci. U.S.A. 69, 3506-3509.
(9) J ensen, O. N., Hojrup, P., and Roepstorff, P. (1991) Plasma
resorption mass spectrometry as a tool in characterization of
abnormal proteins. Application to variant human hemoglobin.
Anal. Biochem. 199, 175-183.
(10) Casida, J . E., Gray, R. A., and Tilles, H. (1974) Thiocarbamate
sulfoxides: Potent, selective and biodegradable herbicides. Science
184, 573-574.
(11) Rossi, R., Barra, D., Bellelli, A., Boumis, G., Canofeni, S.,
Simplicio, P. D., Lusini, L., Pascarella, S., and Amiconi, G. (1998)
Fast-reacting thiols in rat hemoglobins can intercept damaging
species in erythrocytes more efficiently than glutathione. J . Biol.
Chem. 273, 19198-19206.
(12) Rossi, R., Milzani, A., Dalle-Donne, I., Giannerini, F., Giustarini,
D., Lusini, L., Colombo, R., and Simplicio, P. D. (2001) Different
metabolizing ability of thiol reactants in human and rat blood.
J . Biol. Chem. 276, 7004-7010.
(13) Inoue, M., Horiuchi, S., and Morino, Y. (1978) Inactivation of
γ-glutamyl transpeptidase by phenylmethanesulfonyl fluoride, a
specific inactivator of serine enzymes. Biochem. Biophys. Res.
Commun. 82, 1183-1188.
Con clu sion
Molinate covalently modifies rat hemoglobin in vivo in
a cumulative dose dependent manner. Characterization
of the covalent modification demonstrated a S-hexahydro-
1H-azepine-1-carbonyl adduct on the Cys-125 residue of
the â2 and â3 chains consistent with bioactivation of
molinate to a sulfoxide or sulfone intermediate. The
globin modifications produced by molinate are analogous
to those produced by disulfiram, suggesting that molinate
may also covalently modify other proteins that are
modified by disulfiram and share biological effects that
have been demonstrated by disulfiram. Because similar
metabolism to sulfoxide and sulfone metabolites is avail-
able to other thiocarbamate herbicides, the results ob-
tained for molinate may also extend to other compounds
within this class and measurement of cysteine carbamy-
lation may provide a biomarker for assessing exposure
(14) Hart, B. W., and Faiman, M. D. (1992) In vitro and in vivo
inhibition of rat liver aldehyde dehydrogenase by S-methyl N,N-
diethylthiolcarbamate sulfoxide, a new metabolite of disulfiram.
Biochem. Pharmacol. 43, 403-406.
(15) J in, L., Davis, M. R., Hu, P., and Baillie, T., A. (1994) Identifica-
tion of novel glutathione conjugates of disulfiram and diethyl-

Hemoglobin Adduct in Rats Exposed to Molinate
Chem. Res. Toxicol., Vol. 15, No. 2, 2002 217
dithiocarbamate in rat bile by liquid chromatography-tandem
mass spectrometry. Evidence for metabolic activation of disul-
firam in vivo. Chem. Res. Toxicol. 7, 526-533.
(18) Hart, B. W., and Faiman, M. D. (1995) Inhibition of rat liver low
Km aldehyde dehydrogenase by thiocarbamate herbicides. Oc-
cupational implications. Biochem. Pharmacol. 49, 157-163.
(19) Tonkin, E. G., Erve, J . C. L., and Valentine, W. M. (2000)
Disulfiram produces a non-carbon disulfide-dependent Schwan-
nopathy in the rat. J . Neuropathol. Exp. Neurol. 59, 786-797.
(20) Potrepka, R. F., and Morrissey, R. L. (1991) 18-Month Dietary
mouse oncogenicity study with R-4572, unpublished study con-
ducted by Ciba-Geigy Corp., submitted by ICI Americas Inc. to
DPR in Report No. T-13211, DPR Vol. 228-107, No. 096396.
(16) Mays, D. C., Nelson, A. N., Lam-Holt, J . P., Fauq, A. H., and
Lipsky, J . J . (1996) S-methyl N,N-diethylthiocarbamate sulfoxide
and S-methyl N,N-diethylthiocarbamate sulfone, two candidates
for the active metabolite of disulfiram. Alcohol Clin. Exp. Res.
20, 595-600.
(17) Yourick, J . J ., and Faiman, M. D. (1991) Disulfiram metabolism
as a requirement for the inhibition of rat liver mitochondrial low
Km aldehyde dehydrogenase. Biochem. Pharmacol. 42, 1361-
1366.
TX015564A