Development and Application of a Rapid Immunoassay for Difenzoquat in Wheat and Barley Products

Article abstract of DOI:10.1021/jf950290g

A sensitive and simple enzyme-linked immunosorbent assay (ELISA) is described for the quantification of difenzoquat (DFQ) in foods using polyclonal antibodies. Two hapten analogues of DFQ with five-carbon spacer arms attached to one of the aromatic rings were synthesized. The resulting antiserum was specific to DFQ. The minimum detection limit was 0.8 ng/mL for beer and 16 ng/g for cereals with an IC₅₀ (50% inhibition of binding) of 0.28 ng/mL in this assay. The recoveries of DFQ spiked at three levels into beer, cereal, and bread ranged from. 72% to 101%. The mean intraassay and inter-assay coefficients of variation in this procedure were 4.6% and 6.9% for five commodities spiked at 100 ng/g or ng/mL DFQ, respectively. The ELISA procedure was applied to a limited survey of 13 beers and 12 breads, but no detectable DFQ residue was found.

Full text of DOI:10.1021/jf950290g

376
J. Agric. Food Chem. 1996, 44, 376−380
Develop m en t a n d Ap p lica tion of a Ra p id Im m u n oa ssa y for
Difen zoqu a t in Wh ea t a n d Ba r ley P r od u cts
J upiter M. Yeung,* Richard D. Mortimer, and Peter G. Collins
Food Research Division, Food Directorate, Bureau of Chemical Safety, Health Protection Branch,
Health Canada, PL 2203D, Ottawa, Ontario, Canada K1A 0L2
A sensitive and simple enzyme-linked immunosorbent assay (ELISA) is described for the quantifica-
tion of difenzoquat (DFQ) in foods using polyclonal antibodies. Two hapten analogues of DFQ with
five-carbon spacer arms attached to one of the aromatic rings were synthesized. The resulting
antiserum was specific to DFQ. The minimum detection limit was 0.8 ng/mL for beer and 16 ng/g
for cereals with an IC50 (50% inhibition of binding) of 0.28 ng/mL in this assay. The recoveries of
DFQ spiked at three levels into beer, cereal, and bread ranged from 72% to 101%. The mean intra-
assay and inter-assay coefficients of variation in this procedure were 4.6% and 6.9% for five
commodities spiked at 100 ng/g or ng/mL DFQ, respectively. The ELISA procedure was applied to
a limited survey of 13 beers and 12 breads, but no detectable DFQ residue was found.
Keyw or d s: Difenzoquat; herbicide; ELISA; immunoassay; beer; cereal; food
INTRODUCTION
Difenzoquat [1,2-dimethyl-3,5-diphenylpyrazolium me-
thylsulfate (see Figure 1)], is a herbicide that is used
principally to control wild oats in cereal crops. It is not
metabolized by plants or animals and degrades only by
photolysis. Although it is not a heavily used pesticide,
the contribution of its residue in food products to the
total amount of pesticide residues in the Canadian diet
is unknown, in part, because of the inconvenience of its
analysis. The polarity of the difenzoquat molecule
means that it cannot be included in the general screen-
ing method used for the analysis of many pesticides. By
using HPLC (Lawrence et al., 1981), DFQ could be
quantified in wheat products down to 20 ng/g. However,
peaks were broad and asymmetric. Difenzoquat has
also been analyzed by thermospray mass spectrometry
(Barcelo et al., 1993), but this method was not applied
to food products. Most recently, a capillary electro-
phoresis method was developed (Carneiro et al., 1994;
Galceran et al., 1994) for difenzoquat residues but again
not applied to food substances.
Of all the quaternary nitrogen herbicides, paraquat
is the only one for which antibodies have been developed
(Van Emon et al., 1986; Bowles et al., 1992). Paraquat
haptens were prepared by conjugating a pentanoic acid
to one of the pyridinyl nitrogen atoms in the paraquat
molecule. The detection limits reported for the paraquat
immunoassays were 0.1-0.5 ng/mL. Unlike paraquat,
which is dicationic, difenzoquat is monocationic and does
not provide a simple functional group through which it
can be linked to a carrier protein. Conjugation to one
of the pyrazolium nitrogens would mask the cationic
site, which is anticipated to be an important center with
respect to antibody-antigen binding. Our intention,
therefore, was to link a five-carbon spacer arm to one
of the difenzoquat phenyl rings distal to the charged
nitrogen and immunize the rabbits with the resulting
hapten conjugate according to our established procedure
(Newsome et al., 1993).
F igu r e 1. Structure of difenzoquat.
requirement of the cationic site of the hapten for the
recognition of the DFQ epitope was also examined.
MATERIALS AND METHODS
Ch em ica ls a n d Su p p lies. Pesticide standards were ob-
tained from the pesticide repository of the Food Research
Division and were stated to be at least 99% pure by the
respective manufacturers. Cereal, bread, and beer were
purchased in local stores.
Phenylethynyltributyltin, 4-bromobenzoyl chloride, pallad-
ium acetate, triphenylphosphine, methylhydrazine, 4-pen-
tenoic acid, 5% palladium on carbon, platinum oxide mono-
hydrate, and silica gel (200-425 mesh, grade 643, Davisil, 4%
water) for flash chromatography were purchased from Aldrich
Chemical Co. (Milwaukee, WI). Solvents were of distilled-in-
glass grade from Caledon Laboratories (Georgetown, ON).
Methyl iodide was purchased from B.D.H. (Poole, England)
and stored over copper wire. Thin layer chromatography was
done with Whatman MK6F silica plates, 1 in. × 3 in., from
Chromatographic Specialties (Brockville, ON).
Bovine serum albumin (RIA grade), ovalbumin, human
serum albumin, goat anti-rabbit IgG peroxidase conjugate
(second antibody), Tween 20, o-phenylenediamine dihydro-
chloride, 1-ethyl-3-[(dimethylamino)propyl]carbodiimide hy-
drochloride (EDC), and glycerol were purchased from Sigma
Chemical Co. (St. Louis, MO). Freund’s complete and incom-
plete adjuvants were obtained from Gibco (Grand Island, NY).
Dialysis tubing (10 mm in diameter with a 12 000-14 000
molecular weight cutoff) was purchased from Spectrum Medi-
cal Industries Inc. (Los Angeles, CA). Flat-bottom polystyrene
microtiter plates were obtained from Dynatech Laboratories,
Inc. (Chantilly, VA).
Bu ffer s. Phosphate-buffered saline (PBS, pH 7.4), con-
tained 20 mmol of NaH2PO4 and 140 mmol of NaCl per liter
of deionized water. Washing buffer (PBS-T) consisted of 0.1%
Tween 20 in PBS. Coating buffer (pH 9.6) contained 13 mmol
of Na2CO3 and 35 mmol of NaHCO3, while citric buffer (pH
5.0) consisted of 51 mmol of Na2HPO4 and 24 mmol of citric
acid per liter of deionized water. The substrate consisted of
This paper concerns the development of an ELISA
method for the rapid and convenient determination of
difenzoquat residues in food products such as barley,
wheat, and bread (g16 ng/g) and beer (g0.8 ng/mL). The
0021-8561/96/1444-0376$12.00/0
© 1996 American Chemical Society

Rapid Immunoassay for Difenzoquat in Wheat and Barley
J. Agric. Food Chem., Vol. 44, No. 1, 1996 377
5.6 mmol), palladium acetate (4.8 mg, 0.021 mmol) and
triphenylphosphine (13.1 mg, 0.05 mmol) in chloroform (etha-
nol-free) was left at room temperature for 24 h. The reaction
mixture was then diluted with diethyl ether (20 mL) and
stirred for 30 min with potassium fluoride (5 g) in water (10
mL). The precipitated tributyltin fluoride was removed by
filtration and the aqueous phase removed in a separatory
funnel. The remaining organic phase was washed with NaCl-
saturated phosphate buffer (pH 7) before being dried over
anhydrous sodium sulfate, filtered, and evaporated to dryness.
The yellow-brown solid (1.7 g) was dissolved in toluene (12
mL) with pyridinium p-toluenesulfonate (20 mg) and methyl-
hydrazine (0.37 mL, 7.0 mmol) in an Erlenmeyer flask (25 mL)
with a stirring bar and heated on a hot plate (≈100 °C). Beads
of an immiscible liquid (presumably water) were rapidly
formed on the sides of the flask, and TLC analysis (15%
EtOAc-hexane) showed that the reaction was complete within
10 min. The reaction mixture was cooled, diluted with ethyl
acetate, and washed first with water and then with NaCl-
saturated phosphate buffer (pH 7). The organic layer was
dried over anhydrous sodium sulfate, filtered, and evaporated
to dryness to yield a yellow-brown solid (1.68 g). The crude
product was recrystallized from 5% toluene-hexane to yield
light yellow needles, mp 136-137 °C (0.98 g, 57%). TLC (15%
EtOAc-hexane) showed a single spot at Rf 0.3. MS m/z 312,
314 (100%); 1H NMR (CDCl3) δ 3.92 (s, 3H), 6.58 (s, 1H), 7.46
(br s, 5H), 7.52 (d, 2H), 7.70 (d, 2H); 13C NMR (CDCl3) δ 37.61,
103.17, 121.44, 127.05, 128.66, 128.73, 130.46, 131.71, 132.41,
145.27, 149.37.
F igu r e 2. Structure of the linked hapten VII.
Syn th esis of Lin k ed P yr a zole VI. In each of two 4 mL
Reactivials equipped with a magnetic stirring bar were placed
palladium acetate (0.5 mg, 0.002 mmol), triphenylphosphine
(2.5 mg, 0.01 mmol), sodium acetate (79 mg, 1 mmol), 4-pen-
tenoic acid (103 mg, 1 mmol), pyrazole IV (300 mg, 0.96 mmol),
and dimethylformamide (1.5 mL). The Reactivials were then
heated with stirring in an aluminum block at 130 °C for 20 h.
On cooling, the combined mixtures were poured into 1 N NaOH
(15 mL) and extracted with EtOAc (10 mL, 2×). The aqueous
layer was then acidified and extracted with CH2Cl2 (10 mL,
2×). The combined organic layers were dried over sodium
sulfate, filtered, and evaporated to dryness to yield an amber
solid (405 mg). TLC (5% MeOH-CH2Cl2) showed a single
major spot at Rf 0.4, which fluoresced under long UV wave-
length. The crude product was dissolved in 95% EtOH (7 mL)
with PtO2‚H2O (10 mg) and stirred vigorously under a slight
pressure of hydrogen for 3 h (no fluorescent spot remained on
TLC). The reaction mixture was filtered through a 0.45 µm
filter disk and evaporated to dryness to yield a light yellow
solid (399 mg) which showed two or three minor traces on TLC
(5% EtOH-CH2Cl2) besides the major spot at Rf 0.5. Recrys-
tallization from 95% EtOH gave a white solid, mp 124-137
F igu r e 3. Synthetic scheme for the formation of the haptens
VI and VII.
1
°C. MS m/z 334 (100%), 247 (60%); H NMR (CDCl3) δ 1.71
17.5 mg of o-phenylenediamine dihydrochloride (OPD) and 10
µL of 30% H2O2 in 25 mL of citric buffer.
(m, 4H), 2.38 (m, 2H), 2.66 (m, 2H), 3.93 (s, 3H), 6.58 (s, 1H),
7.22 (d, 2H) 7.47 (br s, 5H), 7.73 (d, 2H); 13C NMR (CDCl3) δ
24.37, 30.67, 33.87, 35.29, 37.42, 103.14, 125.60, 128.52,
128.64, 128.68, 128.74, 130.65, 130.93, 141.57, 145.03, 150.56,
178.94.
In str u m en ta tion . Microtiter plates were washed with
PBS-T using a Bio-Rad microplate washer with five wash and
soak cycles for 8 s each. A 12 channel pipetter was used for
dispensing liquids. Optical densities of microtiter wells were
measured on a dual-beam Titertek Multiscan MCC with a 492
nm sample filter and 620 nm reference filters. Data were
transmitted to a spreadsheet program for analysis. The
instrument was checked periodically by a Spectrocheck plate
and software (QC Technology, New York).
Syn th esis of P yr a zoliu m Sa lt VII. A solution of the
pyrazole VI (151 mg, 0.45 mmol) and methyl iodide (0.3 mL,
4.5 mmol) in acetonitrile (2.5 mL) was heated at 100 °C in a
4 mL Reactivial under nitrogen for 5 days. The reaction
mixture was then evaporated to a viscous syrup which was
flash chromatographed on silica gel with acetone to yield an
off-white solid (135 mg). This was recrystallized from acetone-
hexane to give water-soluble, white crystals (102 mg), mp 167-
168 °C (turns yellow on melting). 1H NMR (acetone-d6) δ 1.80
(m, 4H), 2.48 (m, 2H), 2.90 (m, 2H), 4.47 (s, 3H), 4.48 (s, 3H)
7.24 (s, 1H), 7.64 (d, 2H), 7.75-7.94 (br m, 7H); 13C NMR
(acetone-d6) δ 25.20, 31.33, 33.98, 35.92, 36.96, 36.99, 108.43,
124.92, 127.57, 130.21, 130.24, 130.39, 130.42, 132.00, 146.98,
149.37, 149.54, 174.44.
1H and 13C NMR spectra were run on a Bruker 200 MHz
instrument at 22 °C in CDCl3 or acetone-d6 referenced to TMS.
Mass spectra were recorded on a VG Analytical ZAB 2F
instrument in low resolution using a direct inlet probe and
electron impact mode at 70 eV.
SYNTHETIC PROCEDURE
To functionalize DFQ to couple to a protein, a five-carbon
spacer arm was attached to one of the phenyl rings distal to
the pyrazole center (see Figure 2). The synthetic scheme is
outlined in Figure 3.
Syn th esis of P yr a zole IV. A solution of 4-bromobenzoyl
chloride (1.24 g, 5.6 mmol), phenylethynyltributyltin (2.21 g,
Im m u n ogen . The DFQ haptens (VI and VII) were conju-
gated to human serum albumin in the following manner. To
prevent oxidation of the iodide ion, DFQ derivative VII (40
µmol) was dissolved in 0.5 mL of peroxide-free 1,4-dioxane and
0.2 mL of Na2S2O3 (40 µmol) in water. One milliliter of cold

378 J. Agric. Food Chem., Vol. 44, No. 1, 1996
Yeung et al.
0.05 M sodium phosphate (pH 7.0) and 100 mg of EDC were
added to the mixture. After mixing, the reaction mixture was
frozen. Human serum albumin (28 mg) was dissolved in 1.5
mL of the same phosphate buffer and added to the frozen DFQ
mixture. The reaction mixture was placed on an orbital mixer
for 18 h at 4 °C. The resultant conjugate was dialyzed against
running water for 48 h and stored frozen in 1 mL aliquots.
Coa tin g P r otein . Coating proteins were prepared from
ovalbumin by a mixed anhydride coupling reaction rather than
by reaction with EDC to avoid recognition of EDC reaction by
products. The two DFQ haptens, VI and VII (40 µmol), were
each dissolved in 0.5 mL of 1,4-dioxane and 0.2 mL of Na2S2O3
(40 µmol) in water. Fifteen microliters of tri-n-butylamine and
8 µL of isobutyl chloroformate were added. After 60 min in
the dark, the reaction mixtures were frozen. Forty milligrams
of ovalbumin in 3 mL of cold 0.2 M NaHCO3 (pH 9.3) were
added to each of the frozen immunogen mixtures. After 18 h
at 4 °C with agitation, the solutions were dialyzed against
running water for 48 h and aliquots were stored at -20 °C.
The conjugate of immunogen VII coating protein at 1 µg/mL
plus 10 µg/mL ovalbumin was used to sensitize the 96 well
plates, in a manner previously reported (Newsome et al., 1993).
These sensitized plates were used for all of the DFQ analyses.
Im m u n iza tion . Two groups of three, male, New Zealand
White rabbits were used to obtain antibodies against the DFQ
hapten-human serum albumin conjugates. The immunogen
was diluted in PBS and emulsified in Freund’s complete or
incomplete adjuvant to give a concentration of 1 mg/mL.
Rabbits were injected subcutaneously with 0.5 mL of the
complete adjuvant emulsion at four sites. Booster injections
were given at 4 week intervals, substituting incomplete
adjuvant for complete adjuvant.
standard curve for beer. The same beer was used for all of
the beer studies.
RESULTS AND DISCUSSION
As it was anticipated that the development of selective
DFQ antibodies would be enhanced if the linkage to the
albumin protein was most distant from the pyrazolium
charged nitrogen, the synthesis of hapten VII was
pursued (Figure 2). Retrosynthesis of this molecule
suggested that its basic structure could be formed by
the combination of a methylhydrazine moiety and an
appropriately substituted 1,3-diarylpropane unit. The
overall approach is summarized in Figure 3. Logue and
Teng (1982) have demonstrated the construction of
acetylenic ketones from acid chlorides and tributyltin
acetylides, and the formation of the pyrazole ring from
methylhydrazine and an acetylenic ketone is well-
known (Coispeau et al., 1970). Addition of the linker
chain used the convenient conditions of the Heck
reaction (Patel et al., 1977) followed by mild hydrogena-
tion. Although the hydrogenated pyrazole VI had a
wide melting point, neither the 1H nor the 13C NMR
showed any impurity peaks. The possibility of a zwit-
terionic structure in VI in equilibrium with the non-
protonated pyrazole ring may account for the apparent
impurity indicated by the melting point. The most
difficult step was the final methylation reaction which
proved to be very slow.
To test whether the cationic site is critical for the
recognition of the epitope of DFQ, both hapten conju-
gates of VI and VII were used as immunogens. Animals
immunized with the immunogen that lacked the qua-
ternary nitrogen VI produced low titers even after six
booster doses. The IC50, which is the concentration of
the inhibitor that inhibits 50% of the antibody-antigen
binding, ranged from 4 to 226 ng/mL in plates sensitized
with heterologous coating protein VII and required
much higher concentrations of the first antibodies. High
titer but no inhibition by DFQ up to 500 ng/mL was
noted if homologous coating protein VI was used,
suggesting only a small proportion of the IgG would
recognize the epitope of the cationic DFQ. The data
clearly showed that inclusion of the cationic site is
highly preferable but not essential.
The quaternary nitrogen DFQ immunogen VII elic-
ited high titers as early as 1 month which plateaued
after four booster doses. At the dilution of 1:80000,
antiserum showed an IC50 of 0.28 ( 0.06 ng/mL of DFQ
(mean ( SD, n ) 56). The antibodies did not recognize
non-charged DFQ if heterologous coating protein VI was
used. This DFQ specific antiserum also did not cross-
react with any other related pesticides as shown in
Table 1. It is interesting to note that the antibody does,
to a certain extent (0.01%), recognize the common
phenylpropylammonium cationic epitope of cyperquat,
N-methyl-4-phenylpyridinium ion. Examination of the
molecular structure of cyperquat (see Figure 4) shows
that its pendant phenyl group is similarly distant from
its quaternary ammonium center as the phenyl group
in the difenzoquat molecule is from its pyrazolium
centre.
The acid extraction method employed here, for cereals
and breads, was similar to that of Lawrence et al. (1981)
and Van Emon et al. (1987). Unlike the Van Emon
method, we neutralized instead of evaporated the acid.
For consistent results, it is critical to neutralize the acid
with cold, equimolar base prior to the ELISA assay.
Extraction of DFQ in cereal samples with PBS afforded
only 35-45% yield.
Serum titers were monitored. Animals having the highest
titers and most sensitive inhibition curves were exsanguinated
under anaesthesia from 4 to 6 months after the initial
immunization. For long-term storage, the serum was kept
frozen at -20 °C in 100 µL aliquots. Once thawed, an equal
volume of 50% glycerol was added and the solution was stored
at -20 °C. The glycerolated antibodies can be used for at least
6 months with no deleterious effect.
Sa m p le P r ep a r a tion . The commodity to be analyzed
(100-500 g) was thoroughly mixed in a Waring blender.
A
10 g subsample of blended commodity was homogenized in 100
mL of 2 N HCl using a Polytron for 30 s and then sonicated in
an ultrasonic bath for 30 min. Samples were centrifuged at
1500g for 30 min. The solids were removed by filtration
through Whatman No. 1 paper, and about 5-10 mL of filtrate
was collected in polypropylene tubes. J ust prior to analysis,
an equal volume of cold 2 N NaOH was added to a 1 mL aliquot
of the acidic extract while vortexing, and an aliquot (25 µL) of
the neutralized sample was taken for ELISA. The unused
acidic extract was stored in a freezer.
Beer samples were first degassed under vacuum in a
sonicator and then diluted 1:10 with PBS. A 25 µL aliquot
was taken for ELISA.
In the recovery studies, samples were artificially contami-
nated with DFQ at 3 levels (5, 100, and 200 ng/g or ng/mL).
These fortified samples were incubated for 30 min at room
temperature prior to Polytron and ultrasonic extraction.
Im m u n oa ssa y. The ELISA procedure was similar to the
one previously reported (Newsome et al., 1993). Briefly, a 1
mL aliquot of antiserum diluted 1:80000 with 0.1% BSA in
PBS diluent was added to 25 µL of sample or standard. After
mixing and incubation at 4 °C for 60 min, 200 µL was added
to the wells of the sensitized plate in triplicate. After a further
30 min of incubation at 4 °C and washing, a second antibody
horseradish peroxidase conjugate was added. Following fur-
ther incubation at room temperature for 30 min and washing,
the substrate OPD and H2O2 were added. Thirty minutes
later, the color reaction was stopped and the optical densities
were read at 492 nm. The DFQ was determined by a least-
squares plot of the logit of the OD against the log of the
concentration of the standards. The standard curve consisted
of eight concentrations of DFQ (0, 0.02-1.25 ng/mL) in H2O
except for beer. DFQ standards were dissolved in a similarly
diluted (1:10) beer in PBS to provide a matrix-modified

Rapid Immunoassay for Difenzoquat in Wheat and Barley
J. Agric. Food Chem., Vol. 44, No. 1, 1996 379
Ta ble 1. Sp ecificity of An tiser a tow a r d Difen zoqu a t a n d
Oth er Rela ted Com p ou n d s
Ta ble 2. P er cen t Recover ya of Difen zoqu a t fr om Va r iou s
Com m od ities
a
compound
difenzoquat
cyperquat
paraquat
diquat
dodemorph acetate
alachlor
atrazine
IC50 (ng/mL)
difenzoquat added (ng/mL or ng/g)
0.275
2093b
noc
no
no
no
no
no
no
no
n
10
100
200
beerb (lager)
beer (ale)
cereal (barley)
cereal (wheat)
breadc
5
5
3
3
3
87.8 ( 6.3
90.9 ( 4.8
86.1 ( 1.4
79.9 ( 8.5
80.0 ( 4.7
94.2 ( 5.0
100.9 ( 7.7
93.4 ( 10.7
94.2 ( 14.2
98.3 ( 5.1
87.6 ( 8.2
93.7 ( 8.5
76.8 ( 3.8
93.9 ( 13.2
72.1 ( 1.4
a
b
2,4-D
Values are expressed in mean ( SD. Five brands of lager
metalaxyl
triadimefon
captan
and eight brands of ale were tested for DFQ. No detectable DFQ
c
was found. Twelve brands of commercial breads were sampled;
no
no DFQ was detected.
a
IC50 is the concentration of the inhibitor that inhibits 50% of
Ta ble 3. In tr a -a ssa y a n d In ter -a ssa y Coefficien ts of
Va r ia tion of th e P r oced u r e
b
the antibody-antigen binding. At 500 ng/mL cyperquat inhibits
25% binding. c No inhibition at concentrations up to 500 ng/mL.
intra-assay (n ) 3)
inter-assay (n ) 4)
mean ( SD CV (%) mean ( SD CV (%)
beer (lager)
beer (ale)
cereal (barley)
cereal (wheat)
bread
71.4 ( 0.4
88.5 ( 5.2
76.4 ( 6.2
79.6 ( 2.6
90.3 ( 7.4
0.6
5.9
8.1
3.2
8.2
98.6 ( 8.9
79.8 ( 5.1
80.1 ( 4.8
82.6 ( 10.1
94.1 ( 1.0
9.1
6.3
6.0
12.2
1.0
a
Values are expressed in percent recovery when samples were
fortified at 100 ng/g or ng/mL of DFQ.
F igu r e 4. Structure of cyperquat.
parable with the other paraquat immunoassays (Fatori
and Hunter, 1980; Van Emon et al., 1987; Bowles et al.,
1992; Selisker et al., 1995).
To evaluate effectively the sensitivity, reproducibility,
and applicability of the antiserum to different matrices,
it is essential that the analyte in the solution used to
generate the standard curve behaves the same as the
analyte in the sample. Standard curves prepared in
blank bread, barley, or wheat cereals were superimpos-
able on those prepared in water. However, the beer
standard curve was slightly off or tilted from the water
standard curve. The slopes of the standard curves were
not statistically different by Student’s paired t test at p
< 0.01, and the IC50 values were essentially the same.
Bigger daily variations of recoveries were obtained if
water, diluted methanol, or PBS standard curves were
used for beers. As we found that standard curves
generated by either lager or ale were the same, we chose
one light beer to be the matrix modifier for all of the
beer analyses and water for other samples in construct-
ing the standard curves.
The precision of the ELISA was determined by
running the extracts of the samples that were fortified
at 100 ng/g for cereals and bread or 100 ng/mL for beer
three times a day and for 4 days. The intra-assay and
the inter-assay coefficient of variations were 0.6-8.2%
and 1-12.2%, respectively. Results are shown in Table
3. We have applied our procedure for DFQ residues to
13 beer and 12 bread samples purchased in local stores.
No DFQ residues were found, even when beer samples
were run without dilution.
This study demonstrates the feasibility of applying
the ELISA method for the analysis of DFQ residues in
barley and wheat products. The simplicity, sensitivity,
and selectivity of the assay make it suitable for routine
surveillance.
The use of matrix modifier in recent immunoassays
is not uncommon (Silverlight and J ackman, 1994;
Ibrahim et al., 1994a,b; Mountfort et al., 1994). Al-
though we did not demonstrate that the beer matrix
used was DFQ-free, since chromatographic methods
were unable to detect it at our detection range, we did
not use the same beer that was used to construct the
standard curve to do our recovery and inter-assay
coefficients of variation studies. Thus, this approach
indirectly supports the assumption that the beer used
for the standard curve was DFQ-free.
The current tolerance limit for DFQ in wheat and
barley is 0.1 µg/g. Our ELISA was capable of detecting
20 pg/mL DFQ in water based on the lower datum point,
which is at least 10% inhibition, in the standard curve.
Recoveries ranged from 88% to 101% with standard
deviations (SD) <9 at 50, 100, and 200 ng/mL levels of
fortification in beer. However, recoveries for cereals had
a wider spread and ranged from 72% to 98% with SD
e14 at 50, 100, and 200 ng/g levels of spike. Results
are given in Table 2. The minimum detection limit in
the cereal was 16 ng/g. Blank samples were also
assessed, and no false positives were detected. The
recoveries and reproducibilities of the assay were com-
ACKNOWLEDGMENT
We thank D. Bruce Black for the NMR spectra and
Dorcas F. Weber for the mass spectra.
LITERATURE CITED
Barcelo, D.; Durand, G.; Vreeken, R. J . Determination of
quaternary amine pesticides by thermospray mass spec-
trometry. J . Chromatogr. 1993, 647, 271-277.
Bowles, M. R.; Eyles, D. W.; Hampsom, C. G. M.; Pond, S. M.
Quantitation of paraquat in biological samples by radioim-
munoassay using a monoclonal antibody. Fundam. Appl.
Toxicol. 1992, 19, 375-379.
Carneiro, M. C.; Puignou, L.; Galceran, M. T. Comparison of
capillary electrophoresis and reversed-phase ion-pair high-
performance liquid chromatography for the determination
of paraquat, diquat and difenzoquat. J . Chromatogr. 1994,
669A, 217-224.
Coispeau, G.; Elguero, J .; J acquier, R. Action of mono- and
disubstituted-hydrazines on acetylenic ketones. Bull. Soc.
Chim. Fr. 1970, 689-696.
Fatori, D.; Hunter, W. H. Radioimmunoassay for serum
paraquat. Clin. Chim. Acta 1980, 100, 1-90.
Galceran, M. T.; Carneiro, M. C.; Puignou, L. Capillary
electrophoresis of quaternary ammonium ion herbicides:

380 J. Agric. Food Chem., Vol. 44, No. 1, 1996
Yeung et al.
paraquat, diquat and difenzoquat. Chromatographia 1994,
39, 581-586.
vinylic substitution reactions with carboxylic acid deriva-
tives. J . Org. Chem. 1977, 42, 3903-3906.
Ibrahim, A. M. A.; Morsy, M. A.; Hewedi, M. M.; Smith, C. J .
Detection of an aldrin and dieldrin in Egyptian milk samples
using a competitive ELISA. Food Agric. Immunol. 1994a ,
6, 31-38.
Ibrahim, A. M. A.; Ragab, A. A.; Morsy, M. A.; Hewedi, M. M.;
Smith, C. J . Application of an aldrin and dieldrin ELISA to
the detection of pesticides in eggs. Food Agric. Immunol.
1994b, 6, 39-44.
Lawrence, J . F.; Panopio, L. G.; McLeod, H. A. Analysis of
difenzoquat herbicide in wheat products by reversed-phase
liquid chromatography. J . Agric. Food Chem. 1981, 29, 887-
889.
Logue, M. W.; Teng, K. Palladium-catalyzed reactions of acyl
chlorides with (1-alkynyl)tributylstannanes. J . Org. Chem.
1982, 47, 2549-2553.
Selisker, M. Y.; Herzog, D. P.; Erber, R. D.; Fleeker, J . R.; Itak,
J . A. Determination of paraquat in fruits and vegetables by
a magnetic particle based enzyme-linked immunosorbent
assay. J . Agric. Food Chem. 1995, 43, 544-547.
Silverlight, J . J .; J ackman R. Enzyme immunoassay for the
detection of levamisole in meat and milk. Analyst 1994, 119,
2705-2706.
Van Emon, J .; Hammock, B.; Seiber, J . N. Enzyme-linked
immunosorbent assay for paraquat and its application to
exposure analysis. Anal. Chem. 1986, 58, 1866-1873.
Van Emon, J .; Seiber, J .; Hammock, B. Application of an
enzyme-linked immunosorbent assay (ELISA) to determine
paraquat residues in milk, beef, and potatoes. Bull. Environ.
Contam. Toxicol. 1987, 39, 490-497.
Mountfort, K. A.; Reynolds, S. L.; Thorpe, S. A.; White, S. N.
Comparison of ELISA and HPLC techniques for the analysis
of carbendazim and thiabendazole residues in fruit and
vegetables. Food Agric. Immunol. 1994, 6, 17-22.
Newsome, W. H.; Yeung, J . M.; Collins, P. G. Development of
enzyme immunoassay for captan and its degradation prod-
uct tetrahydrophthalimide in foods. J . AOAC Int. 1993, 76,
381-386.
Received for review May 12, 1995. Revised manuscript
received October 11, 1995. Accepted October 23, 1995.X
J F950290G
X Abstract published in Advance ACS Abstracts, De-
cember 1, 1995.
Patel, B. A.; Ziegler, C. B.; Cortese, N. A.; Plevyak, J . E.;
Zebovitz, T. C.; Terpko, M.; Heck, R. F. Palladium-catalyzed