Immunoassay for Simazine and Atrazine with Low Cross-Reactivity for Propazine

Article abstract of DOI:10.1021/jf950583+

An antibody for simazine and atrazine has been developed that exhibits low cross-reactivity to propazine relative to most atrazine antibodies heretofore evaluated. The cross-reactivities obtained in an enzyme-linked immunosorbent assay were 100 ± 4% for simazine, 76 ± 9% for atrazine, and 12.6 ± 1.3% for propazine. This was achieved by immunizing rabbits with the hapten 6-[[[4-chloro-6-(methylamino)]-1,3,5-triazin-2-yl]amino]hexanoic acid coupled to keyhole limpet hemocyanin. The influence of tracer hapten structure on the assay sensitivity was investigated in two competitive formats. The performance of the assay with respect to pH differences and ionic strength was also examined. The lowest IC₅₀ values achieved for simazine were in the 0.1 μg/L range, with the limit of quantitation being 50 ng/L. Spike-recovery studies in tap and ground water as well as analysis of crude ground water samples show the usefulness of this sensitive antibody for simazine detection.

Full text of DOI:10.1021/jf950583+

2210
J. Agric. Food Chem. 1996, 44, 2210−2219
Im m u n oa ssa y for Sim a zin e a n d Atr a zin e w ith Low Cr oss-Rea ctivity
for P r op a zin e
Monika Wortberg, Marvin H. Goodrow, Shirley J . Gee, and Bruce D. Hammock*
Departments of Entomology and Environmental Toxicology, University of California,
Davis, California 95616
An antibody for simazine and atrazine has been developed that exhibits low cross-reactivity to
propazine relative to most atrazine antibodies heretofore evaluated. The cross-reactivities obtained
in an enzyme-linked immunosorbent assay were 100 ( 4% for simazine, 76 ( 9% for atrazine, and
1
6
2.6 ( 1.3% for propazine. This was achieved by immunizing rabbits with the hapten 6-[[[4-chloro-
-(methylamino)]-1,3,5-triazin-2-yl]amino]hexanoic acid coupled to keyhole limpet hemocyanin. The
influence of tracer hapten structure on the assay sensitivity was investigated in two competitive
formats. The performance of the assay with respect to pH differences and ionic strength was also
examined. The lowest IC50 values achieved for simazine were in the 0.1 µg/L range, with the limit
of quantitation being 50 ng/L. Spike-recovery studies in tap and ground water as well as analysis
of crude ground water samples show the usefulness of this sensitive antibody for simazine detection.
Keyw or d s: Immunoassay; enzyme-linked immunosorbent assay (ELISA); simazine; atrazine;
propazine; triazine herbicide
INTRODUCTION
Triazine herbicides are used extensively for control-
ling broad-leaf weeds in agriculture or for road main-
tenance. Atrazine is the most extensively used pesticide
in the United States (U.S. EPA, 1994). Simazine and
atrazine are among the herbicides that have been
detected in ground water in several areas of the United
States for many years (Ritter, 1990; Aharonson et. al.,
F igu r e 1. Structures of the three triazine herbicides simazine,
atrazine, and propazine.
1
987). In California, where approximately 20% of all
pesticides used in the United States were applied as of
986, simazine is used more extensively than atrazine
ity to atrazine, and 136% cross-reactivity to propazine.
This was similar to the findings of Schlaeppi et al.
1
(
1989): 100% for atrazine, 90% for propazine, and 2.5%
for simazine. Another monoclonal antibody described
by Karu et al. (1991) was more specific for propazine
(
Domagalski and Dubrovsky, 1992).
Due to a growing need to monitor pesticide contami-
nation, several immunoassays for atrazine have been
developed as screening tools for environmental water
and soil samples (Bushway et al., 1988; Wittmann and
Hock, 1989; Schlaeppi et al., 1989; Karu et al., 1991;
Harrison et al., 1991; Giersch, 1993). To immunochem-
ically quantify simazine instead of atrazine, atrazine
antibodies have been used (Goh et al., 1992; Lucas et
al., 1991). However, most of these antibodies lack the
ability to detect simazine sensitively as well. Only one
antibody has been designed to detect simazine, but it
detects atrazine equally well (Eremin, 1995). Most
atrazine antibodies cross-react with simazine approxi-
mately 5-20% when atrazine is used as the standard
(
196%) than for atrazine (100%) or for simazine (31%).
This monoclonal antibody was raised by immuniza-
tion with a conjugate derived from 3-[[4-(ethylamino)-
6
-[(1-methylethyl)amino]-1,3,5-triazin-2-yl]thio]propano-
ic acid.
As immunoassays are used increasingly to distinguish
among analytes in a single class of compounds, there is
an increasing need for antibodies of high specificity. The
key to the cross-reactivity characteristics lies in the
immunizing hapten and, therefore, our goal was to find
a better tailored immunogen to generate antibodies with
higher binding to simazine than to atrazine or pro-
pazine. By introducing the smallest possible alkylamino
group, methylamine, into the immunizing hapten side
chain (Figure 2), we envisioned the generation of
antibodies with at least enhanced selectivity for si-
mazine. Our anticipation was that a small methyl-
amine group would generate a small antibody pocket
incapable of accommodating the much larger steric
requirement of the isopropyl moieties of atrazine and
propazine and hence provide selectivity for the ethyl-
amino groups of simazine. This paper presents the
results of this approach to produce simazine specific
antibodies.
(100%). In addition, antibodies produced specifically
against atrazine usually exhibit a much higher cross-
reactivity to propazine, which is the main analyte for
these antibodies. This has been discussed previously
by Giersch (1993). The structures of the three herbi-
cides are shown in Figure 1.
Eremin (1995) obtained antibodies displaying equiva-
lent cross-reactivity for both atrazine and simazine and
3
6
1% for propazine when immunizing with 6-[[[4-chloro-
-(ethylamino)]-1,3,5-triazin-2-yl]amino]hexanoic acid.
Harrison et al. (1991), however, using the same immu-
nogen, developed polyclonal antibodies that exhibited
4
0% cross-reactivity to simazine, 100% cross-reactivity
MATERIALS AND METHODS
to atrazine, and 72% cross-reactivity to propazine.
Giersch (1993) obtained a monoclonal antibody with
only 4% cross-reactivity to simazine, 100% cross-reactiv-
Ch em ica ls. All chemicals used for the hapten synthesis,
dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) of
S0021-8561(95)00583-8 CCC: $12.00
© 1996 American Chemical Society

Immunoassay for Simazine and Atrazine
J. Agric. Food Chem., Vol. 44, No. 8, 1996 2211
R
f
0.27 [hexane-ethyl acetate (2:1 v/v) plus 2% acetic acid],
0
3
1
7
6
.81 [acetonitrile-water-acetic acid (88:12:2 v/v/v)]; IR (KBr)
267 (m, NH), 3121 (w, NH), 1701 (m, CdO), 1572 (vs, CdN),
-
1 1
228 (m, C-O) cm ; H NMR (DMSO-d
.8 (m, 1 H, CH NH), 7.6 (m, 1 H, CH NH), 3.2 (m, 2 H, CH
), 2.76, 2.72 (two d, J ) 4.6 Hz, 3 H, CH ), 2.19 (t, J ) 7.4
-2), 1.5 (m, 4 H, CH -3,5), 1.3 (m, 2 H, CH -4)
O the 12.0, 7.8, and 7.6 ppm peaks disappeared,
the 3.2 ppm multiplet became a triplet at 3.27 ppm with J )
.6 Hz, and the two doublets centered at 2.76 and 2.72 ppm
6
) δ 12.0 (s, 1 H, OH),
2
3
2
-
3
Hz, 2 H, CH
with added D
2
2
2
(
2
6
1
3
became two singlets); C NMR (DMSO-d
68.5, 168.4, 167.7 (Ar, C-4), 166.3, 166.1, 165.9 (Ar, C-2 or
6), 165.8, 165.7, 165.2 (Ar, C-2 or -6), 40.3, 40.2 (C-6), 34.0
C-2), 29.0, 28.7 (C-5), 27.6, 27.4 (CH ), 26.2 (C-4), 24.5 (C-3);
6
) δ 174.9 (COOH),
1
-
(
3
F igu r e 2. Synthetic route for the immunizing hapten.
+
FAB-MS, m/z (relative intensity) 276 (35, M + H + 2), 275
+
+
(
26, M + H +1), 274 (100, M + H ); FAB-HRMS calcd for
274.1071, obsd 274.1049.
Im m u n ogen Syn th esis. Hapten III was coupled to KLH
liquid chromatography (LC) grade, tetramethylsilane (TMS),
N-hydroxysuccinimide (NHS), and dicyclohexylcarbodiimide
10 5 2
C H16ClN O
(
DCC) were obtained from Aldrich (Milwaukee, WI). Silica
gel 60 F254 plastic-backed thin layer chromatography plates
of 0.25-mm thickness were purchased from E. Merck (Darm-
stadt, Germany). Triazine herbicide standards were obtained
from Ciba-Geigy (Greensboro, NC). Horseradish peroxidase
via a modified activated NHS ester method using a water
soluble carbodiimide (Tijssen, 1985). For coupling, a 20 mg/
mL stock solution of NHS in DMF, a 5 mg/mL stock solution
of EDC in DMF, and a 10 mg/mL hapten stock solution in DMF
were prepared. Five hundred microliters (17 µmol) of the
hapten solution was mixed with 794 µL (21 µmol) of the EDC
solution and 159 µL (28 µmol) of the NHS stock and stirred
for 7 h at room temperature. Fifty milligrams of 65% pure
KLH was dissolved in a 1:1 mixture of PBS buffer (phosphate-
(
HRP) conjugates of goat anti-rabbit IgG as well as goat anti-
rabbit IgG, ovalbumin (ova) grade VI, crude ovalbumin,
-ethyl-3-[3-(dimethylamino)propyl]carbodiimide (EDC), and
1
tetramethylbenzidine (TMB) were purchased from Sigma
Chemical Co. (St. Louis, MO). Keyhole limpet hemocyanin
(KLH) of 65% purity was obtained from Calbiochem (La J olla,
buffered saline) and 0.13 M NaHCO
3
to a final concentration
CA) and HRP from Boehringer (Mannheim, Germany). Buffer
reagents of analytical grade were purchased from Fisher
Scientific (Fair Lawn, NJ ). For purification of hapten-protein
conjugates, 5 and 10 mL Presto desalting columns (Pierce,
Rockford, IL) were used. Microtiter plates were obtained from
Nunc (Roskilde, Denmark).
of 4 mg/mL and cooled to 4 °C. The organic solution was added
in 50 µL aliquots to 5 mL of the protein solution under constant
stirring; 0.75 mg (2.6 µmol) of hapten/mg of KLH was used,
which equals addition of a total volume of 1226 µL of the
organic solution to 5 mL of the aqueous solution. The solution
was stirred for 30 min at 4 °C and then overnight at room
temperature. The turbid solution was then centrifuged to
remove precipitated protein. Subsequently, the KLH-conju-
gate was passed through gel filtration columns equilibrated
in PBS and the fractions containing the conjugate were pooled.
The protein content of the pooled column fractions was tested
by spectrophotometry (E280-E260, Kalckar method) and deter-
mined to approximately 2 mg/mL. Note that the pooled
column fractions precipitated shortly after they were collected.
Therefore, it is recommended to apply the centrifuged super-
natant immediately to the column to prevent formation of a
precipitate in the column.
Im m u n iza tion P r otocol. Three 6-month-old female New
Zealand White rabbits (no. 2281, 2282, and 2283) were
immunized every 4 weeks and bled 10 days after each
injection. The immunogen for the three rabbits consisted of
200 µL of the KLH-hapten III conjugate (400 µg of protein)
diluted in 1.6 mL of PBS mixed with 1.6 mL Freund’s
adjuvant. For the initial immunization Freund’s complete
adjuvant was used; for all subsequent immunizations the
incomplete adjuvant was used. The components were emulsi-
fied, and 1 mL of the mixture was injected into each rabbit
subcutaneously to deliver 125 µg of KLH per rabbit. For the
first three bleeds, 5 mL of blood was collected each time, bleeds
four and five provided 15 mL each. Rabbit 2282 was exsan-
guinated during the fifth bleed.
Appar atu s. Melting points were determined with a Thomas-
Hoover apparatus (A. H. Thomas Co., Philadelphia, PA) and
are uncorrected. Infrared (IR) spectra were recorded on a
Mattson Galaxy Series FTIR 3000 spectrometer (Madison, WI).
1
13
H- and C-NMR spectra were measured on a General Electric
QE-300 spectrometer (Bruker NMR, Billerica, MA) operating
at 300.1 and 75.5 MHz, respectively. Chemical shifts (δ) are
expressed in parts per million downfield from internal TMS.
Fast atom bombardment low- and high-resolution mass spectra
(FAB-MS and -HRMS) were obtained on a ZAB-HS-2F spec-
trometer (VG Analytical, Wythenshawe, U.K.) using xenon (8
keV, 1 mA) for ionization and 3-nitrobenzyl alcohol as the
matrix. Polyethylene glycol 300 was added as a mass cali-
brant. Optical densities were read on a Molecular Devices
UVMax reader (Sunnyvale, CA) equipped with the standard
ELISA software Softmax (Molecular Devices). The software
package for fitting sigmoidal calibration curves and for cal-
culating concentrations of samples was based on the log-
logistic four-parameter fit.
H a p t en Syn t h esis: 6-[[[4-Ch lor o-6-(m et h yla m in o)]-
1
,3,5-tr ia zin -2-yl]a m in o]h exa n oic Acid (III) (F igu r e 2).
To a vigorously stirred solution of 7.38 g (40.0 mmol) of
cyanuric chloride (I) in 100 mL of acetone maintained at -5
to -3 °C was added dropwise over 0.5 h a solution of 2.70 g
(
40.0 mmol) of methylamine hydrochloride in 10 mL of water.
This was followed by the dropwise addition of a slurry of 6.72
g (80.0 mmol) of NaHCO
and 20 mL of water over 0.5 h at
Coa tin g Ha p ten Syn th esis. Ovalbumin conjugates of a
variety of triazine haptens were synthesized according to an
adapted standard NHS ester method as described elsewhere
(Wortberg et al., 1995). For synthesis of coating hapten
conjugates the hapten to protein ratio was as low as 5:1 or
10:1 to ensure a sufficiently low avidity desirable for competi-
tive immunoassay.
3
the same temperature range. After an additional 1 h of
stirring at this temperature, the mixture was allowed to warm
slowly to 15 °C. Then 5.51 g (42.0 mmol) of solid 6-amino-
hexanoic acid was added followed by a slurry of 7.06 g (84.0
3
mmol) of NaHCO in 20 mL of water. After stirring overnight
at room temperature, 30% NaOH was added until the pH
reached 11. The mixture was filtered through Celite, and the
residue was thoroughly washed with 200 mL of water. The
filtrate and washes were combined and rotoevaporated to
remove the acetone and then acidified to pH 1 with 6 M HCl.
The resultant precipitate, a pasty, white solid, was collected
and dried in a vacuum desiccator over Drierite to constant
weight, 3.73 g (34%, crude yield), mp 146.5-148.0 °C (dec).
Compounds were detected on TLC first by viewing under UV
light (254 nm) and then by staining in an iodine chamber. TLC
Tr a cer Syn th esis w ith Hor ser a d ish P er oxid a se. HRP
tracers were synthesized according to a modified version of
Schneider and Hammock (1991) using the NHS ester method
via DCC. A 35:1 molar ratio of hapten to enzyme was used.
Stock solutions in dry DMF containing 10 mg/mL hapten were
prepared. For synthesis of XIV-HRP, as an example, 72 µL
(0.72 mg, 2.4 µmol) of the hapten XIV stock solution was mixed
with 32.5 µL (6 µmol) of a 20 mg/mL solution of NHS in DMF.
Then 3.1 mg (12 µmol) of DCC was added. The organic
mixture was stirred for 20 h at room temperature and

2212 J. Agric. Food Chem., Vol. 44, No. 8, 1996
Wortberg et al.
Both the H- and 13C-NMR spectra determined at room
1
centrifuged to separate the precipitated urea. Two milligrams
-
2
(
4.4 ×10 µmol) of HRP was dissolved in 3 mL of 0.13 M
NaHCO and cooled to 4 °C. Of the organic solution, 73 µL
35-fold molar excess) was added in 10 µL aliquots to the
temperature displayed multiple peaks for those atoms
in or adjacent to the ring. At 80 °C the number of H
3
1
(
peaks collapsed to that expected for one structure except
for the N-methyl doublet of doublets, which appeared
as two singlets. At ambient temperature, a correlated
spectroscopy (COSY) spectrum permitted definitive
assignments of each set of peaks and was especially
useful for assigning the two N-H peaks. The lower field
(7.8 ppm) multiplet displayed coupling to the N-CH2
appendage (3.2 ppm), whereas the higher field (7.6 ppm)
multiplet was coupled to the N-CH3 (2.76, 2.72 ppm)
stirred HRP solution over several minutes. The solution
became turbid and a precipitate formed. After stirring over-
night at 4 °C, precipitated protein was removed by centrifuga-
tion. The supernatant was passed through a gel filtration
column equilibrated with PBS. Fractions containing the HRP
conjugate were pooled and frozen.
Coa tin g Ha p ten F or m a t En zym e-Lin k ed Im m u n osor -
ben t Assa y (ELISA). Optimal concentrations of the hapten-
ovalbumin conjugates were determined by two-dimensional
titration. A working dilution of the conjugates of 1:10000 in
PBS was used, unless otherwise indicated. The coating hapten
solution (150 µL) was incubated in wells of microtiter plates
overnight at 4 °C. Alternatively, coating was performed at
room temperature over a 2 h incubation period. The wells
were then emptied and incubated for 30 min with 150 µL of a
13
peaks. A C distortionless enhanced by polarization
transfer (DEPT) experiment confirmed the assignment
of the 27.6 and 27.4 ppm peaks to the N-CH3 moiety
as well as substantiating the presence of the 40.3 and
0.2 ppm C-6 peaks masked, but surmised due to the
unsymmetrical appearance of the DMSO solvent peaks
4
0
.5% (w/v) ova-PBS solution (blocking solution) without
in the normal C-NMR spectrum. Finally, 1H- C
heteronuclear correlation spectroscopy (HETCOR) con-
firmed the assignment of the four protons at 1.5 ppm
to the C-3 (24.5 ppm) and C-5 (29.0, 28.7 ppm) carbon
atoms and the two protons at 1.3 ppm to the C-4 (26.2
ppm) carbon atom of the hexanoic acid side chain.
1
3
13
intermediate rinsing. The plates were then washed four times
with PBS containing 0.05% Tween 20 (PBST) and were ready
to use or were stored for up to 5 weeks at 4 °C in PBS
containing 0.02% NaN
3
. Triazine stock solutions were pre-
pared in DMSO (1.00 mg/mL). Starting from the organic stock
solution, triazine standards were diluted with PBS or other
buffers, tap water, or ground water. The optimal antibody titer
of serum 2282 was determined to be 1:10000 in a two-
dimensional titration. For competition, 50 µL of diluted
antiserum solution and 100 µL of sample or standard were
incubated together for 1 h. For all assays, samples and
standards were pipetted in triplicates. The wells were washed
four times with PBST and then incubated with 100 µL of a
goat anti-rabbit IgG-HRP conjugate (diluted 1:12500 in PBS)
for 1 h. After four washing steps, 100 µL of substrate solution
was added (2.4 mg of TMB dissolved in 400 µL of DMSO plus
An tiser a Scr een in g. Initially, the titer of each bleed
was tested in a two-dimensional titration against a
homologous system; i.e., the immunizing hapten coupled
to ovalbumin instead of KLH was used as a coating
hapten. This ensured that increase in titer after
subsequent boost injections was due to specific recogni-
tion of the hapten and not to binding to the KLH moiety
of the conjugate. Incorporation of blanks using ova as
coating alone showed that the binding was specific to
the hapten. After the third bleed (second boost), no
significant increase of the titers was observed for any
of the three antisera.
1
00 µL of 1% H
pH 5.5). The reaction was stopped after 15-30 min by adding
0 µL of 2 M H SO
Tr a cer F or m a t ELISA. Wells were coated by incubating
with 100 µL of goat anti-rabbit IgG diluted 1:2000 in PBS at
°C overnight. The wells were then washed four times with
PBST and incubated for 1 h with 100 µL of specific antibody
282, diluted 1:10000 in PBS. The wells were emptied with-
2 2
O in 25 mL of 0.1 M sodium acetate buffer of
5
2
4
.
The affinity of the antisera to heterologous systems
structurally different coating haptens) was qualita-
4
(
tively tested in a comparative binding study, using the
second bleed. For this experiment coating haptens were
diluted 1:1500, the antisera 1:2000. Coating haptens
exhibiting a sufficiently high antibody affinity (OD
>0.2-0.3) but a lower affinity than the immunogen (OD
<0.9) were further investigated. We have found that
coating haptens that bind with titers more like the
immunogen result in assays that are poorly inhibited
by free analyte (Goodrow et al., 1995). The triazine
haptens used in this study and the results of the
qualitative binding assessment are presented in Table
2
out a washing step, and 150 µL of the ovalbumin blocking
solution was added. After 30 min, the plates were rinsed four
times and were then used immediately or stored in PBS
containing 0.02% NaN . For competition studies, 100 µL of
3
standard or sample was incubated together with 50 µL of the
HRP tracer (in triplicates). Optimal HRP tracer dilutions were
determined to be 1:1000-1:2000 (in PBS) by two-dimensional
titration. Standards were pipetted first, and then the tracer
was added. After 1 h of incubation, the wells were rinsed four
times and subsequently incubated with 100 µL of substrate
solution. The substrate conversion was stopped by adding 50
1
. These coating haptens were then tested in a com-
2 4
µL of 2 M H SO .
petitive assay using a series of simazine standards and
a suitable antibody concentration of later bleeds. Quali-
tative examination of coating hapten binding with later
bleeds revealed no major differences to the data shown
here.
RESULTS AND DISCUSSION
Ha p ten Syn th esis. The immunizing hapten, 6-[[4-
chloro-6-(methylamino)]-1,3,5-triazin-2-yl]amino]hex-
anoic acid (III) (Figure 2), was synthesized by treating
cyanuric chloride (I) in acetone with 1 equiv of methyl-
amine hydrochloride and 2 equiv of aqueous NaHCO3
while maintaining the reaction mixture at -5 to -3 °C.
One equivalent of solid 6-aminohexanoic acid was then
added followed by 2 equiv of aqueous NaHCO3 at room
temperature to replace the second chlorine of II. Acidi-
fication provided crude hapten III, which required
recrystallizations from three different solvents to obtain
a pure product.
After IV-ova was chosen as a suitable coating hapten
for all antisera and optimizing dilutions, cross-reactivi-
ties of the third bleeds to an array of triazine herbicides
were tested in a preliminary fashion. As was predicted
from the structure of the immunizing hapten, all three
antisera showed lower cross-reactivities to propazine
than to atrazine or simazine. Antiserum 2281 showed
a preference for atrazine (100%) with a high simazine
cross-reactivity (68%). The slope of all calibration
curves was low, however (B ranged from 0.5 to 0.75).
Antiserum 2283 responded equally well to simazine and
atrazine (both 100%) with an acceptable slope of 0.7-
0.9, whereas antiserum 2282 clearly favored simazine
1
Structural confirmation was accomplished by H and
C NMR, HRMS, MS, and IR. All of these were
consistent with the assigned structure for the product.
1
3

Immunoassay for Simazine and Atrazine
J. Agric. Food Chem., Vol. 44, No. 8, 1996 2213
Ta ble 1. Tr ia zin e Ha p ten s Used for Coa tin g Ha p ten or HRP Tr a cer Syn th esisa
R1
N
N
2
NR3
H
R N
N
H
no.
III
IV
V
R1
R2
CH3
R3
2281
2282
2283
ref
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
(CH2)5-COOH
(CH2)-COOH
(CH2)2-COOH
(CH2)3-COOH
(CH2)4-COOH
(CH2)5-COOH
(CH2)-COOH
(CH2)2-COOH
(CH2)5-COOH
(CH2)5-COOH
(CH2)5-COOH
(CH2)5-COOH
(CH2)5-COOH
(CH2)2-COOH
CH2CH3
+++
+
+++
+++
+++
+++
nt
+++
+
+++
+++
+++
+++
nt
+++
+
+++
+++
++
+++
nt
CH2CH3
CH2CH3
CH2CH3
CH2CH3
CH2CH3
CH(CH3)2
CH(CH3)2
CH(CH3)2
H
CH(CH3)2
CH(CH3)2
H
CH2CH3
CH(CH3)2
CH2CH3
CH(CH3)2
H
Goodrow et al., 1990
Goodrow et al., 1990
Goodrow et al., 1990
Goodrow et al., 1990
Goodrow et al., 1990
Goodrow et al., 1990
Goodrow et al., 1990
Goodrow et al., 1990
Lucas et al., 1995
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
XIX
XX
nt
nt
nt
Cl
Cl
+++
+++
-
++
0
+++
+++
-
++
0
+++
+++
0
OCH3
SCH3
OH
OH
SOCH3
Kido et al., 1996
Kido et al., 1996
++
+
Goodrow, unpublished
Lucas et al., 1993a
Huber and Hock, 1986
Lucas et al., 1993b
Goodrow et al., 1990
Muldoon et al., 1994
++
+
+
+
+
-
-
N-acetylcysteine
S(CH2) 2-COOH
S(C6H6)-COOH
CH(CH3)2
CH2CH3
-
-
++
+
+
0
H
-
-
a
The degree of binding of the antisera 2281-3 to these haptens when used as coating haptens is indicated qualitatively, using the
absolute OD as a measure of affinity. A KLH-III conjugate was the immunogen. Key: - ) OD <0.2; 0 ) OD 0.2-0.4; + ) OD 0.4-0.6;
+
+ ) OD 0.6-0.8; +++ ) OD >0.8; nt, not tested in the initial screen.
(
100%) over atrazine (83%) and had slope parameters
defined as cross-reactivity >0.2%) to antibody 2282
(fourth bleed) are shown in Figure 4. Since all cross-
reactivities are relative to the simazine curves, the IC50
and the standard error for simazine were determined
from six replicate curves to give a statistically useful
standard deviation. All other IC50 values and the
resulting cross-reactivities were determined in repli-
cates of four curves each. The errors of the cross-
reactivities were calculated by adding the individual
CVs for the averaged IC50 values of simazine and the
respective cross-reactant.
B between 0.7 and 1.1. Further studies were conducted
with 2282 only since it had the highest preference for
simazine and a sufficient slope. Bleed five of antibody
2
282 was somewhat more cross-reactive to atrazine than
bleed four, while showing no change in sensitivity for
simazine; thus, bleed four was chosen for optimization.
Although 2283 may be of practical use as a generic
triazine antibody since it detects atrazine and simazine
equally well, the goal of this study was the development
of a simazine selective antibody. If screening for atra-
zine and simazine was desired, 2283 would be preferred.
Since we are interested in distinguishing structurally
related compounds from each other (Wortberg et al.,
As described above, the antiserum 2282 showed lower
cross-reactivity to propazine than to atrazine or si-
mazine. Virtually all atrazine antibodies reported in
the literature to date are actually propazine antibodies,
a fact that is usually ignored because this compound is
not a commonly used herbicide. Most research groups
have designed atrazine immunogens by using an ethyl-
amino group or an isopropylamino group and a five-
carbon handle terminated with a functional group,
which mimics an “extended” ethylamino group. The
result usually is an antibody that detects propazine best
(two isopropyl groups), followed by atrazine (one iso-
propyl group) and simazine (no isopropyl substituent).
The affinity is observed to be higher for an analyte that
has one carbon atom more than the immunogen or is
more branched than the immunogen at the same
position. In this study, we maintained the handle
location and length but reduced the size of the remain-
ing alkylamino group by one carbon to methylamine.
The result is an antibody that again favorably binds
analytes with one carbon more than the immunogen.
Consequently, we obtained antibodies that detect si-
mazine and atrazine very well, but none is specific for
simazine. Since the methylamine substituent of the
immunogen is distinctly smaller than an isopropyl
group, the isopropyl groups of propazine do not fit very
well into the binding pocket of the antibody. As a
1
996) and/or quantitation of mixtures (J ones et al., 1994;
Wortberg et al., 1995) by using an array of antibodies
with differing specificities and multivariate statistics,
the “simazine” antibody 2282 was used.
After the preliminary experiments showed that an-
tibody 2282 was the antibody displaying the most
desirable properties, it was further tested against all
hapten conjugates that seemed useful according to the
qualitative binding experiment. All coating haptens
were then diluted 1:7500, the antiserum was diluted
1
:10000, and the simazine standard concentrations were
chosen according to the location of the curve midpoints.
These midpoints were estimated in an initial assay. The
dose-response curves for simazine against an array of
different coating haptens are shown in Figure 3. The
IC50 values obtained with different hapten structures
cover a range of almost 4 orders of magnitude, thus
showing that hapten structure is very crucial for assay
sensitivity. The best haptens have extremely short
handles with just one carbon atom between the triazine
ring and the functional group. Additionally, they have
a Cl atom and the ethylamino or an isopropylamino
group.
Cr oss-Rea ctivity Stu d ies. A list of the triazines
tested for cross-reactivity and their respective struc-
tures is given in Table 2. The cross-reactivities of
those triazine herbicides that bind significantly (here
2
consequence of the ethylamine moiety in position R (see
Table 2), deisopropylatrazine has a relatively high cross-
reactivity. Deethylatrazine has a lower cross-reactivity

2214 J. Agric. Food Chem., Vol. 44, No. 8, 1996
Wortberg et al.
metabolites, 2- or 3-fold dealkylated metabolites, and
triazine precursors. Terbutryn with its methylthio
substituent cross-reacts only negligibly (<0.2%), but
terbutylazine exhibits a cross-reactivity of as much as
1
1%. Surprisingly, there is no significant difference in
cross-reactivity between propazine and terbutylazine.
One would expect terbutylazine to cross-react less than
propazine because of its even bulkier tertiary butyl
substituent. But in contrast to propazine, terbutylazine
also bears one less bulky ethylamino group instead of
only isopropylamino groups. These findings indicate
2
that besides the alkylamino group in position R , the
Cl atom plays a key role in the recognition process. If
both of the features are present, the third substituent
on the triazine has less influence on the cross-reactivity.
As a consequence of this preference, antibody 2282 is
class-specific for chloro-s-triazines and exhibits a high
selectivity for simazine and atrazine.
En zym e Tr a cer Assa ys. Figure 5 shows an array
of simazine calibration curves obtained with different
HRP tracers. Note that the range of curve midpoints
is only spread out over 2 orders of magnitude as opposed
to almost 4 orders of magnitude in the coating hapten
format. Several different approaches to couple hapten
IV to HRP (all based on variations of NHS ester
activations) did not produce a tracer that exhibited any
affinity to the antibody, although all showed high
enzyme activity. The coupling ratio was not confirmed,
but when we used the same batch of hapten-NHS ester
in a parallel experiment to synthesize an ovalbumin
conjugate, we obtained a functioning coating hapten.
That indicates that the triazine hapten was sufficiently
activated for coupling. Also, earlier coupling experi-
ments with a similar triazine derivative (XI), which has
a five-carbon spacer instead of a one-carbon spacer,
yielded an HRP tracer with sufficient affinity to the
antibody as well as a functioning ovalbumin-based
coating hapten. This indicates that generally coupling
of NHS-activated carboxylic acid triazine derivatives to
HRP is a functioning method.
Our conclusion is that the conjugation of IV-HRP
was most likely successful, but the tracer would not bind
to the antibody. It appears likely that the optimal
spacer length for an HRP tracer is different from the
optimal spacer length for an ovalbumin conjugate. We
speculate that in case of the unsuccessful tracer with
the short handle, the entire triazine moiety merely
“disappeared” inside some pocket in the enzyme, thus
rendering antibody binding sterically hindered. This
possibility is currently under investigation.
The same effect would then explain another interest-
ing finding. Hapten VIII with its long five-carbon
handle yielded a very insensitive assay in the coating
hapten format (Figure 3A), supposedly due to the
antibodies’ high affinity to the conjugate (quasi-homolo-
gous hapten), but when the same hapten was used as
an HRP tracer, the affinity was lower than expected
(Figure 5A), which was indicated by the relatively low
curve midpoint of the tracer format assay. Still, this
hapten resulted in one of the “poorest” tracer format
assay compared to haptens with shorter handles.
Com p a r ison of t h e Tw o F or m a t s. Since spacer
length is apparently more crucial in the coating hapten
format than in the tracer format assay, with a very short
length yielding the best assays, we must assume that
the concept of an “apparent handle length” should be
considered. With the HRP tracer, differences in handle
length seem to be less pronounced, as can be seen in
F igu r e 3. Dose-response curves for simazine using an array
of different coating haptens. n ) number of carbon atoms in
3
the handle in R ; Et, ethylamino; iPr, isopropylamino; OH,
1
hydroxy; MPA, mercaptopropanoic acid in R . The structure
of the hapten largely determines the location of the IC50 and
therefore the assay sensitivity. Haptens with short spacers
(“handles”) generally result in more sensitive assays. No curves
were obtained with haptens XIII, XV, XVII, XVIII, and XX.
The homologous system III was not tested. Three replicates
of each standard were analyzed. The simazine concentrations
for each calibration curve were chosen after the approximate
location of the IC50 was known from previous experiments.
than deisopropylatrazine because its remaining side
chain is an isopropyl group. Triazines with a methoxy
or a methylthio substituent exhibit very little or no
cross-reactivity at all; the same is true for hydroxy

Immunoassay for Simazine and Atrazine
J. Agric. Food Chem., Vol. 44, No. 8, 1996 2215
Ta ble 2. Str u ctu r e of th e Tr ia zin e Com p ou n d s Tested for Cr oss-Rea ctivity
R1
N
N
R2
N
R3
triazine herbicide
atrazine
cyanazine
procyazine
propazine
simazine
terbutylazine
ametryn
prometryn
simetryn
terbutryn
prometon
terbumeton
hydroxyatrazine
hydroxysimazine
deethylatrazine
deisopropylatrazine
didealkylated atrazine
cyanuric acid
ammelide
R1
R2
R3
Cl
Cl
Cl
Cl
Cl
Cl
SCH3
SCH3
SCH3
SCH3
OCH3
OCH3
OH
OH
Cl
Cl
Cl
NHCH2CH3
NHCH2CH3
NHcyclopropyl
NHCH(CH3)2
NHCH2CH3
NHCH2CH3
NHCH2CH3
NHCH(CH3)2
NHCH2CH3
NHCH2CH3
NHCH(CH3)2
NHCH2CH3
NHCH2CH3
NHCH2CH3
NH2
NHCH(CH3)2
NHCCN(CH3)2
NHCCN(CH3)2
NHCH(CH3)2
NHCH2CH3
NHC(CH3)3
NHCH(CH3)2
NHCH(CH3)2
NHCH2CH3
NHC(CH3)3
NHCH(CH3)2
NHC(CH3)3
NHCH(CH3)2
NHCH2CH3
NHCH(CH3)2
NHCH2CH3
NH2
NH2
NH2
OH
OH
NH2
NH2
OH
OH
OH
OH
NH2
NH2
NH2
NH2
ammeline
melamine
atrazine mercapturate
N-acetylcysteine
NHCH2CH3
NHCH(CH3)2
assays despite using a homologous tracer. One would
assume that using a homologous hapten system in an
assay for a small molecule automatically results in an
insensitive assay. This was indeed observed in our
study with ovalbumin conjugates of the immunogen III
(
data not shown). Investigators using the homologous
system either used monoclonal antibodies (Giersch,
993; Schlaeppi et al., 1989) that had been screened to
1
result in a sensitive assay or polyclonal antibodies in
combination with an HRP tracer (Wittmann and Hock,
1
989; Eremin, 1995). There is no evidence in the
literature of a successful polyclonal (triazine) assay
employing a homologous albumin conjugate.
For quantitative comparison of curve midpoints the
exact hapten density has to be determined, e.g. by
matrix-assisted laser desorption/ionization mass spec-
trometry (MALDI-MS) as was demonstrated by Wen-
gatz et al. (1992). Coupling ratios may differ from
protein to protein even when using the same excess of
reagents for coupling, and different coupling densities
lead to different affinities (avidity concept) and thus
different IC50 values.
In flu en ce of Ion ic Str en gth a n d p H on th e
Assa y. To evaluate the variability of the coating hapten
format assay due to matrix effects, we prepared si-
mazine standards in several different buffer systems.
One approach was to vary the pH, another was changing
the ionic strength but maintaining constant pH. We
also investigated the influence of an organic solvent. For
the pH variation study we used several 0.01 M buffers
ranging from pH 4.0 to 10.0, which all contained 0.125
M NaCl. The NaCl served as an “ion buffer” to level
out differences between buffer systems and to adjust the
total ionic strength to the physiological saline concen-
tration of 0.135 M. Buffers of pH 4.0, 5.0, 6.0, 7.0, and
8.0 were prepared by adjusting the pH of 0.01 M NaH2-
PO4 and Na2HPO4 solutions with 1 M NaOH and 1 M
HCl, respectively. The buffer of pH 9.0 contained Tris-
HCl; the buffer of pH 10.0 was made from glycine-
NaOH. The effect of pH variation on the calibration
curve with antibody 2282 is shown in Figure 6A.
F igu r e 4. Cross-reactivity pattern of antibody 2282 against
a panel of triazine herbicides. The error bars represent the
standard deviation calculated from replicate calibration curves
which were obtained with the same set of standards (n ) 6
curves for simazine, n ) 4 for all others). All other triazines
listed in Table 1 showed a cross-reactivity of less than 0.2%.
Figure 5 compared to Figure 3. This indicates that the
apparent handle length of a given hapten is different
in both assay formats and is not necessarily constant
as one would assume from the triazine hapten structure
alone. Thus, discussion of the optimal handle length
should be modified by discussing an apparent handle
length, which may differ from protein to protein. This
idea would help explain the findings of Schneider and
Hammock (1991) and Eremin (1995), from which it was
observed that in a heterologous system a shorter handle
yielded a better assay in an HRP tracer format, but in
an example with an alkaline phosphatase (AP) tracer,
the reverse appeared to be true (Schneider and Ham-
mock, 1991). The idea that a hapten behaves differently
in terms of affinity when coupled to different proteins
may also explain why some groups developed sensitive

2216 J. Agric. Food Chem., Vol. 44, No. 8, 1996
Wortberg et al.
A
B
F igu r e 5. Dose-response curves for simazine using an array
of different HRP-hapten tracers. Again, the handle length
determines the assay sensitivity. In the tracer format differ-
ences in IC50 values are less pronounced with the same haptens
than in the coating hapten format. No binding was observed
to haptens IV and IX. All other haptens not shown have not
been investigated. (For abbreviations, see Figure 3.)
F igu r e 6. (A) Influence of varying pH on the simazine cali-
bration curve. The ionic strength was constant in all buffers.
(
B) Influence of the ionic strength on the simazine calibration
curve. Low ionic strength results in higher optical density but
also in higher IC50
To evaluate the changes in curve parameters due to
ionic strength variation, we used 10 times concentrated
PBS, PBS, and 10 times diluted PBS, all pH 7.5. In
addition, we spiked a batch of PBS buffer with 10%
methanol and another batch with 0.5 M Na2SO4. We
also used tap water and ground water for comparison.
Calibration curves obtained with simazine standards in
dilutions of different ionic strength are shown in Figure
.
The curves generated in phosphate buffers pH 5.0-
8.0 show little variation. At pH 4.0 the OD drops
slightly, at pH 10.0, it decreases very sharply. Curve
quality in terms of correlation coefficient and sufficient
ODmax (>0.5) deteriorates at pH 9.0 and above as well
as at pH 4.0. The IC50 values obtained at pH 5.0-8.0
max
6
B.

Immunoassay for Simazine and Atrazine
J. Agric. Food Chem., Vol. 44, No. 8, 1996 2217
range from 0.13 to 0.26 µg/L, but there is no correlation
observed between the direction of the pH change and
the increase or decrease of the IC50. In general, it seems
important that the sample pH is between 5.0 and 8.0-
9
.0 to ensure good curve quality. However, to level out
differences between samples of different pH values, one
could easily use a PBS buffer with a higher buffering
capacity for diluting the antiserum. Also, one does not
expect to observe extreme pH differences in drinking
water samples. The pH range of the ground water
samples (see below) was only from 6.9 to 7.7. In general,
we observed an increase in IC50 with a decrease in ionic
strength. The nature of the ions present is not as
important as the ion concentration (more specifically,
ionic strength). The highest IC50 was obtained with tap
water (0.35 µg/L) and ground water (0.40 µg/L), the
lowest (0.15 µg/L) with PBS. High ionic strengths
reduced the overall binding and resulted in lower
maximum optical densities. The ODmax dropped from
0
.8 to 0.4 when the buffer concentration was increased
from 1 times PBS (0.135 M solution) to 10 times PBS
1.35 M solution) or when the PBS buffer contained 0.5
(
M Na2SO4. Both resulted in less precise curves in terms
of curve fit (Figure 6B). These observations are in
qualitative agreement with the earlier findings of Lucas
et al. (1991) and Muldoon and Nelson (1994). They also
indicate that the ion concentration of the matrix is
crucial for the accurate analysis of real samples. When
environmental samples are analyzed, the matrix effect
has to be minimized by diluting the sample, “buffered
out” by adding a sufficiently concentrated buffer (ionic
strength adjustment), or the standards must be pre-
pared in a blank matrix that is as similar to the environ-
mental samples as possible. Since ground water samples,
as well as tap water, may have a lower ionic strength
than PBS buffer, as was observed herein, the overall
ionic strength in the assay should be increased by
preparing antibody or tracer solution in more concen-
trated PBS. This will ensure an ionic buffering as well
as a pH effect and also will result in a low IC50. We
used 3 or 10 times PBS for the following experiments.
F igu r e 7. Spike recovery for simazine in tap water at various
concentration levels. The regression data for the recovery study
2
are R ) 0.943, b ) 0.994, and a ) 0.041 (52 points).
Ta ble 3. Sp ik e Recover ies in Gr ou n d Wa ter for both th e
Coa tin g Ha p ten a n d th e Tr a cer F or m a ta
spike level
µg/L)
recovery
(µg/L)
std error
(µg/L)
recovery
(%)
std error
(%)
(
IV-ovalbumin
0
0.007
0.037
0.113
0.531
0.006
0.004
0.025
0.039
0
0
0
.05
.1
.5
74
113
106
9.8
22
7.4
XI-HRP
0
0.015
0.069
0.107
0.398
0.004
0.010
0.006
0.041
0
0
0
.05
.1
.5
138
107
80
15
5
10
a
n ) 6 samples were tested at each concentration level.
An a lysis of Gr ou n d Wa ter Sa m p les. We obtained
8 ground water samples from the California Depart-
An a lysis of Sp ik ed Wa ter Sa m p les. To examine
the spike recovery, tap water (source: Davis, CA) and
ground water (source: San J oaquin Valley, CA) samples
were spiked with different levels of simazine. In one
study, tap water was spiked over the entire range of
the assay from 0.05 to 0.5 µg/L, using four replicates at
each spike level (individually prepared). The results are
shown in Figure 7. The regression data for the recovery
2
ment of Pesticide Regulation, CalEPA. The samples
were collected throughout the San J oaquin Valley of
California in the fall of 1994 and had been analyzed for
simazine and a few other herbicides. Data were ob-
tained by GC; positives were confirmed by HPLC.
ELISA analyses were conducted without prior knowl-
edge of the GC results.
Initially, raw ground water samples were analyzed
by using calibration curves in PBS to identify negatives.
Then we repeated the assay using a standard curve in
a blank (negative) matrix, adding reagents in 3 times
PBS buffer. The results obtained from a coating hapten
format ELISA were correlated to the data provided by
CalEPA as plotted in Figure 8.
The ELISA results compared favorably with the GC
data. Since the immunoassay tends to slightly overes-
timate concentrations, as can be seen from spike-
recovery data, false positives are likely to occur. Since
the limit of quantitation was defined as 50 ng/L, all
concentrations found below that level were set auto-
matically to zero; this corresponds to the practice of the
reference GC method. Since these GC negative samples
are not necessarily simazine-free, there can be a bias
between both methods. The immunoassay might esti-
mate a concentration greater than the limit of its
quantitation, whereas the GC analysis still detects it
2
study are R ) 0.943, b ) 0.994, and a ) 0.041 (52 data
points).
In a second study, tap and ground water samples were
each spiked with four concentration levels (0, 0.05, 0.1,
and 0.5 µg/L) and analyzed. To not simply evaluate the
repetition error of the assay, the six replicate samples
of each concentration were prepared individually. The
data are shown in Table 3. Spike-recoveries in both
assay formats were compared. The recoveries at the
lowest simazine concentration of 50 ng/L, which we
define as the limit of quantitation, were 74 ( 10% in
the coating hapten format and 138 ( 15% in the tracer
format. The best recoveries, as expected, were around
the IC50 concentration of 0.1 µg/L (113 ( 22% and 107
(
5%, respectively), because here the curve has the
highest slope. Since different ground or tap waters
might result in different recoveries, these data are
interpreted as a general indicator of the performance
of the assay in ground and tap water.

2218 J. Agric. Food Chem., Vol. 44, No. 8, 1996
Wortberg et al.
and the GC data. We thank A. D. J ones of the Facility
of Advanced Instrumentation (UCD) for mass spectra,
the Department of Chemistry (UCD) for the use of their
IR and NMR instruments, P. V. Choudary of the
Antibody Engineering Laboratory (UCD) for providing
his research facility, and J . Chandler for technical
support. B.D.H. is a Burroughs Wellcome Toxicology
Scholar, M.W. is a fellow of the Deutsche Forschungs-
gemeinschaft.
LITERATURE CITED
Aharonson, N.; Cohen, S. Z.; Drescher, N.; Gish, T. J .; Gorbach,
S.; Kearney, P. C.; Otto, S.; Roberts, T. R.; Vonk, J . W.
Potential contamination of ground water by pesticides. Pure
Appl. Chem. 1987, 10, 1419-1446.
Bushway, R. J .; Perkins, B.; Savage, S. A.; Lekousi, S. J .;
Ferguson, B. S. Determination of atrazine residues in water
and soil by enzyme immunoassay. Bull. Environ. Contam.
Toxicol. 1988, 40, 647-654.
Domagalski, J . L.; Dubrovsky, N. M. Pesticide residues in
ground water of the San J oaquin Valley, California. J . Hydr.
1
992, 130, 299-338.
Eremin, S. A. Polarization fluoroimmunoassay for rapid,
specific detection of pesticides. In Immunoanalysis of Agro-
chemicals; eds. Nelson, J . O., Karu, A. E., Wong, R. B., Eds.;
ACS Symposium Series 586; American Chemical Society:
Washington, DC, 1995; pp 223-234.
F igu r e 8. Analysis of ground water samples. Simazine
detected by ELISA vs simazine measured by GC. Regression
2
data are R ) 0.674, b ) 1.321, and a ) 0.024 (28 points).
Giersch, T. A new monoclonal antibody for the sensitive detec-
tion of atrazine with immunoassay in microtiter plate and
dipstick format. J . Agric. Food Chem. 1993, 41, 1006-1011.
Goh, K. S.; Spurlock, F.; Lucas, A. D.; Kollmann, W.; Schoenig,
S.; Braun, A. L.; Stoddard, P.; Biggar, J . W.; Karu, A. E.;
Hammock, B. D. Enzyme-linked immunosorbent assay
as below that limit. Thus, a “false positive” is inevitably
generated. However, no false negatives were observed
and only one sample gave a strong false positive (0.2
µg/L instead of <0.05 µg/L).
Immunoassay data were generated using untreated,
crude samples, whereas the GC analyses required time-
consuming and costly solid-phase or liquid-liquid ex-
traction procedures. Thus, the immunoassay developed
here is a means of quickly and inexpensively assessing
the degree of water contamination by simazine.
Con clu sion s. The development of an antibody with
highest cross-reactivity toward the triazine herbicide
simazine was achieved by choosing a tailored immuno-
gen with a methylamine substituent on the triazine ring
instead of an ethylamine substituent. Implementation
of this antibody in our multianalyte immunoassay
system (MELISA; J ones et al., 1994; Wortberg et al.,
(ELISA) of simazine for Delhi and Yolo soils in California.
Bull. Environ. Contam. Toxicol. 1992, 49, 348-353.
Goodrow, M. H., Harrison R. O.; Hammock, B. D. Hapten
synthesis, antibody development, and competitive inhibition
enzyme immunoassay for s-triazine-herbicides. J . Agric.
Food Chem. 1990, 38, 990-996.
Goodrow, M. H.; Sanborn, J . R.; Stoutamire, D. W.; Gee, S. J .;
Hammock, B. D. Strategies for immunoassay hapten design.
In Immunoanalysis of Agrochemicals; Nelson, J . O., Karu,
A. E., Wong, R. B., Eds.; ACS Symposium Series 586;
American Chemical Society: Washington, DC, 1995; pp
1
19-139.
Harrison, R. O.; Goodrow, M. H.; Hammock, B. D. Competitive
inhibition ELISA for the s-triazine herbicides: assay opti-
mization and antibody characterization. J . Agric. Food
Chem. 1991, 39, 122-128.
1
995) should improve the performance by better dis-
criminating simazine from atrazine than in the initial
MELISA system.
Huber, S. J .; Hock, B. Atrazine in water. In Methods of
Enzymatic Analysis, Vol. XII, Drugs and Pesticides; Berg-
meyer, H. U., Ed.; VCH: Weinheim, 1986.
J ones, G.; Wortberg, M.; Kreissig, S. B.; Bunch, D. S.; Gee, S.
J .; Hammock, B. D.; Rocke, D. M. Extension of the four-
parameter logistic model for ELISA to multianalyte analy-
sis. J . Immunol. Methods 1994, 177, 1-7.
Karu, A. E.; Harrison, R. O.; Schmidt, D. J .; Clarkson, C. E.;
Grassman, J .; Goodrow, M. H.; Lucas, A. D.; Hammock, B.
D.; Van Emon, J . M.; White, R. J . Monoclonal immunoassay
of triazine herbicides. ACS Symp. Ser. 1991, No. 451,
Another means to obtaining an antibody with a higher
simazine over atrazine selectivity may be the use of an
immunogen with one or two methylamine groups with
1
attachment to the protein through the R -position. The
1
most common handle used at R has been the mercap-
toalkanoic acid. The resulting antibody, however, would
likely be mainly selective for simetryn and possibly
ametryn, whereas simazine and atrazine would only be
minor analytes as predicted from the report of Harrison
et al. (1991). If, however, the presence of simetryn and
ametryn is not expected in a sample, this approach
might be one solution to the atrazine-simazine prob-
lem. An antibody of such specificity would prove very
valuable in discriminating subclasses of triazine herbi-
cides in the MELISA system. To achieve generation of
an antibody with specificity for simazine remains a
challenge since the structures of atrazine and simazine
are very closely related.
5
9-77.
Kido, H.; Goodrow, M. H.; Griffeth, V.; Lucas, A. D.; Gee, S.
J .; Hammock, B. D. Development of an enzyme-linked
immunosorbent assay (ELISA) for the detection of hydroxy-
triazines. J . Agric. Food Chem. 1996, submittted for pub-
lication.
Lucas, A. D.; Schneider, P.; Harrison, R. O.; Seiber, J . N.;
Hammock, B. D.; Biggar, J . W.; Rolston, D. J . Determination
of atrazine and simazine in water and soil using polyclonal
and monoclonal antibodies in enzyme-linked immunosor-
bent assay. Food Agric. Immunol. 1991, 3, 155-167.
Lucas, A. D.; Bekheit, H. K. M.; Goodrow, M. H.; J ones, D. A.;
Kullman, S; Matsumura, F.; Woodrow, J . E.; Seiber, J . N.;
Hammock, B. D. Development of antibodies against hy-
ACKNOWLEDGMENT
We thank Kean Goh, Department of Pesticide Regu-
lation, CalEPA, for providing the ground water samples

Immunoassay for Simazine and Atrazine
J. Agric. Food Chem., Vol. 44, No. 8, 1996 2219
droxyatrazine and hydroxysimazine: application to environ-
mental samples. J . Agric. Food Chem. 1993a, 41, 1523-1529.
Lucas, A. D.; J ones, D. A.; Goodrow, M. H.; Saiz, S. G.; Blewett,
C.; Seiber, J . N.; Hammock, B. D. Determination of atrazine
metabolites in human urine: development of a biomarker
of exposure. Chem. Res. Toxicol. 1993b, 6, 107-116.
Wengatz, I.; Schmid, R. D.; Kreissig, S.; Wittmann, C.; Hock,
B. Determination of the hapten density of immunoconju-
gates by matrix-assisted UV laser desorption/ionization
mass spectrometry. Anal. Lett. 1992, 25, 1983-1997.
Wittmann, C.; Hock, B. Improved enzyme-immunoassay for
the analysis of s-triazines in water samples. Food Agric.
Immunol. 1989, 1, 211-224.
Wortberg, M.; Kreissig, S. B.; J ones, G.; Rocke, D. M.; Ham-
mock, B. D. An immunoarray for the simultaneous deter-
mination of multiple triazine herbicides. Anal. Chim. Acta
Lucas, A. D.; Goodrow, M. H.; Seiber, J . N.; Hammock, B. D.
Development of an enzyme-linked immunosorbent assay for
the N-dealkylated s-triazines: application to environmental
and biological samples. Food Agric. Immunol. 1995, 7, 227-
2
41.
1
995, 304, 339-352.
Muldoon, M. T.; Huang, R. N.; Hapeman, C. J .; Fries, G. F.;
Ma, M. C.; Nelson, J . O. Hapten synthesis and immunoassay
development for the analysis of chlorodiamino-s-triazine in
treated pesticide waste and rinsate. J . Agric. Food Chem.
Wortberg, M.; J ones, G.; Kreissig, S. B.; Rocke, D. M.; Gee, S.
J .; Hammock, B. D. An approach to the construction of
an immunoarray for differentiating and quantitating
cross reacting analytes. Anal. Chim. Acta 1996, 319, 291-
1
994, 42, 747-755.
3
03.
Muldoon, M. T.; Nelson, J . O. Evaluation of the effects of
selected agricultural materials on an ELISA for s-triazines.
Food Agric. Immunol. 1994, 6, 357-370.
Received for review August 24, 1995. Revised manuscript
Ritter, W. F. Pesticide contamination of ground water in the
United Statessa review. Environ. Sci. Health 1990, B25,
X
received April 4, 1996. Accepted April 29, 1996. This
research was funded by NIEHS Superfund 2P42-ES04699,
USDA Forest Service NAPIAP G-5-95-20-062, U.S. EPA
Center for Ecological Health Research CR 819658, NIEHS
Center for Environmental Health Sciences 1P30-ES05707, and
Water Resources Center W-840. Although the information in
this document has been funded in part by the U.S. Environ-
mental Protection Agency, it may not necessarily reflect the
views of the Agency and no official endorsement should be
inferred.
1
-29.
Schlaeppi, J .-M.; F o¨ ry, W.; Ramsteiner, K. Hydroxyatrazine
and atrazine determination in soil and water by enzyme-
linked immunosorbent assay using specific monoclonal
antibodies. J . Agric. Food Chem. 1989, 37, 1532-1538.
Schneider, P.; Hammock, B. D. Influence of the ELISA format
and the hapten-enzyme conjugate on the sensitivity of an
immunoassay for s-triazine herbicides using monoclonal
antibodies. J . Agric. Food Chem. 1991, 40, 525-530.
Tijssen, P. Conjugation of haptens. In Practice and Theory of
Enzyme Immunoassay; Elsevier: Amsterdam, 1985.
U.S. EPA. Pesticide industry sales and usage, 1992 and 1993
market estimates; 733-K-94-001; U.S. government Printing
Office, Washington, DC, J une 1994.
J F950583+
X
Abstract published in Advance ACS Abstracts, J une
15, 1996.