pubs.acs.org/JAFCPublished on Web 04/13/2010
©2010 American Chemical Society
J. Agric. Food Chem. 2010, 58, 5241–5247
Monoclonal Antibody-Based Enzyme-Linked Immunosorbent
Assays for the Organophosphorus Insecticide O-Ethyl
O-4-Nitrophenyl Phenylphosphonothioate (EPN)
JEE YOUN SHIM,†YOUNG AH KIM,†YONG TAE LEE,‡BRUCE D. HAMMOCK,§AND
†Department of Food Science and Nutrition, Kyungpook National University, Daegu 702-701, Korea,
‡Department ofMolecular Life Science, YeungnamUniversity, Gyongsan 712-749, Korea, and§Department
of Entomology and Cancer Research Center, University of California, Davis, California 95616
This study aimed at developing competitive direct and indirect enzyme-linked immunosorbent
assays (ELISAs) for the organophosphorus insecticide O-ethyl O-4-nitrophenyl phenylphosphono-
thioate (EPN) using a monoclonal antibody (mAb). Of the five EPN derivatives (haptens) prepared
for use as an immunogen or as a competitor, two of them were used as the immunogen for the
production of the mAbs. By using the antibody with the highest specificity and a coating antigen
(hapten-OVA conjugate), a competitive indirect ELISA was developed, which showed an IC50of 2.9
ng/mL with a detection limit of 0.3 ng/mL. A competitive direct ELISA using a different antibody and
an enzyme tracer was also developed, which showed an IC50of 0.6 ng/mL with a detection limit of
0.09 ng/mL. The mAbs in both assays showed negligible cross-reactivity with other organopho-
sphorus pesticides. The recoveries of EPN from spiked samples determined by the developed
ELISA ranged from 59 to 143%. Dilution of the samples improved the recovery. The assay
performance of the present ELISAs based on the mAb was compared with that of the EPN ELISAs
based on polyclonal antibodies (pAbs) that had been developed previously and was found to be
better in dynamic response.
KEYWORDS: EPN; insecticide; immunoassay; monoclonal antibody; ELISA
Analytical methods involving gas chromatography (GC) and
high-performance liquid chromatography (HPLC) have been
these classical methods require a high cost, skilled analysts, and
time-consuming sample preparation steps. Therefore, there is a
growing demand for more rapid and economical methods for
determining pesticide residues. Immunoassays began recently to
gain acceptance as an alternative to the traditional methods that
can meet such demands as they are fast, sensitive, and cost-
effective tools for detecting trace amounts of chemicals such as
O-Ethyl O-4-nitrophenyl phenylphosphonothioate (EPN) is
range of insects (3). The most sensitive and toxicologically
relevant effect after the administration of EPN is the inhibition
of acetylcholinesterase activity (4). EPN is also reported to be an
endocrine-disrupting chemical with estrogenic and antiandro-
genic activity (5). Analysis of EPN is carried out by multiresidue
methods using GC (6).
This paper describes the development of mAb-based ELISAs
for EPN. An ELISA for this pesticide based on pAbs has been
developed previously by the current authors (7), but an ELISA
based on a mAb has not yet been reported. Although mAbs are
difficult to prepare due to a complicated and time-consuming
procedure, they are increasingly used in immunoassays because
they have the merit that their use makes available virtually
unlimited amounts of specific antibodies. However, mAbs tend
to be of lower affinity than their corresponding counterparts by
about an order of magnitude (8). Recently, we have developed
several pesticide ELISAs based on both pAbs and mAbs (9-12)
and have observed certain differences in assay performance
the possibility of using mAbs for merits other than those men-
Whereas a pAb is a heterogeneous mixture of antibody mole-
cules, a mAb is homogeneous. Therefore, mAbs are presumed to
be more specific than pAbs. Because this assumption appears not
to have been verified, we attempted to test it in this study on an
experimental basis. Another postulation we considered to be
plausible on the basis of our comparison of the standard curves
of pAb- and mAb-based ELISAs for the same pesticides was that
method. Therefore, the additional merits of using mAbs that we
attempt to identify are assay performance with regard to assay
specificity and dynamic response. We tested this hypothesis by
comparing the assay performance of mAb-based and pAb-based
in previous studies (9-12). Comparison of the assay specificity
5242 J. Agric. Food Chem., Vol. 58, No. 9, 2010 Shim et al.
was evaluated on the basis of the cross-reactivity of antibodies.
Dynamic response, also called “calibration sensitivity”, was then
evaluated by investigating the standard curves with regard to the
slope of the curve and the clarity of plateau formation.
MATERIALS AND METHODS
Chemicals and Instruments. Organophosphorus pesticides including
EPN were purchased from Dr. Ehrenstorfer (Augsburg, Germany).
Horseradish peroxidase(HRP)-labeled goatanti-mouseIgG and sorbitan
monolaurate (Tween 20) were obtained from Sigma (St. Louis, MO).
Tetramethylbenzidine (TMB) was obtained from Boehringer Mannheim
(Mannheim, Germany). Microtiter plates (Maxisorp) were purchased
from Nunc (Roskilde, Denmark). ELISA plates were washed with a
1575 Immuno-wash from Bio-Rad (Hercules, CA), and well absorbances
were read with a Vmax microplate reader obtained from Molecular
Devices (Menlo Park, CA). GC measurements were made using a Clarus
600 gas chromatograph with an electron capture detector (ECD) from
Perkin-Elmer (Waltham, MA).
Synthesis of Haptens and Hapten-Protein Conjugates. The
haptens used for immunization, antigen coating, and enzyme labeling
are described in Figure 1. The synthetic procedures for these haptens have
been previously reported (7). Hapten A was attached to keyhole limpet
hemocyanin (KLH) to be used as the immunogen. Haptens A-E were
attached to ovalbumin (OVA) for use as coating antigens. Haptens A-E
were conjugated with HRP for use as enzyme tracers. The conjugation
method used was the active ester method (7,9). Two hapten/HRP molar
ratios (10 and 50) were employed in the synthesis of enzyme tracers.
the hapten-KLH conjugates, that is, immunization, cell fusion, hybrido-
papers (10). The immunogens used were hapten A-KLH and hapten
B-KLH conjugates. Procedures performed differently were as follows.
noncompetitive indirect ELISAs for hybridoma screening. Ascites rather
ascites, female BALB/c mice were given a “priming” intraperitoneal
injection of 0.5 mL of tetramethylpentadecane (pristine oil). Ten days
hybridoma cells. Ascites fluids were harvested by peritoneal tap with an
18-gauge needle on the seventh day after the cells were introduced,
followed by purification by ammonium sulfate precipitation.
Competitive Indirect ELISA. Checkerboard assays, in which anti-
bodies were titrated against various amounts of the coating antigen, were
used to measure the reactivity of antibodies and to select an appropriate
procedure for the checkerboard assays was the same as that for compe-
titive assays (see below) except that addition of the pesticide standard or
sample is omitted at the competition step.
From the results of the checkerboard assays, three antibodies were
selected as the most suitable ones (6A, 4A, and 8B). Then, to select the
most suitable coating antigen, competitive assays were performed under
various combinations of immunoreagents at several concentration levels.
The concentrations of the antibodies and the coating antigen chosen were
further optimized. Additionally, thetolerance ofELISA tovarious water-
miscible organic solvents used to dissolve pesticides was tested for assay
optimization. For this test, standard pesticide solutions were prepared in
various concentrations of acetone, acetonitrile, or methanol (10, 20, 40,
and 80% in PBS, which became 5, 10, 20 and 40%, respectively, after
combination with equal volumes of diluted antibody). The effect of
buffering capacity of the assay solution on ELISA performance was also
studied using different concentrations of phosphate ion in 20% metha-
nol-PBS to dissolve the pesticide (10, 90, 190, and 390 mM phosphate,
which became 10, 50, 100, and 200 mM, respectively, after combination
with equal volumes of diluted antibody). The influence of pH of the assay
solution was also studied.
The procedure of the competitive assay was as follows. All incubations
except that for antigen coating were carried out at room temperature
(25 ?C) and, after eachincubation, the plates werewashedfour times with
were coated with hapten-OVA (200-1000 ng/mL, 100 μL/well) in
carbonate-bicarbonate buffer (50 mM, pH 9.6) by overnight incubation
at 4 ?C. The plates were blocked by incubation with 1% gelatin in PBS
(200 μL/well) for 1 h. Serial dilutions of the analyte in methanol-PBS
(50 μL/well) were added, followed by 50 μL/well of antibody diluted
(1/1000-1/10000) with PBS. After incubation for 1 h, 100 μL/well of a
diluted (1/6000) goat anti-mouse IgG-HRP was added. The mixture was
incubated for 1 h, and 100 μL/well of a TMB solution (400 μL of 0.6%
TMB-DMSO and 100 μL of 1% H2O2diluted with 25 mL of citrate-
acetate buffer, pH 5.5) was added. The reaction was stopped after an
absorbance was read at 450 nm.
Competitive Direct ELISA. A checkerboard assay, in which anti-
used to optimize the amount of antibody and enzyme tracer. The
procedure for the checkerboard assays was the same as that for
competitive assays (see below) except that addition of pesticide stan-
dard or sample is omitted at the competition step. After selection of the
most suitable antibody and enzyme tracer from the checkerboard
The influence of kinds and concentrations of organic solvents, phos-
phate ion concentration, and pH of the assay buffer on ELISA per-
formance was also studied using the same procedure as that for the
The assay procedure was as follows. All incubations except that for
precoating the wells with IgG were carried out at room temperature
(25 ?C) and, after each incubation, plates were washed four times with
PBST. Microtiter plates were precoated with anti-mouse IgG (5 μg/mL,
100 μL/well) in carbonate-bicarbonate buffer (50 mM, pH 9.6) by
overnight incubation at 4 ?C. The plates were loaded with 100 μL/well
well of a diluted (1/100-1/5000) enzyme tracer in 10 mM PBS. After
absorbance was read at 450 nm.
Figure 1. Structures of haptens for EPN.
ArticleJ. Agric. Food Chem., Vol. 58, No. 9, 2010 5243
Determination of Cross-Reactivities. Several organophosphorus
ELISA procedure described above. The cross-reactivity values were
calculated as follows: (IC50of EPN/IC50of compound) ? 100.
samples were obtained from a local food supplier. Solutions of EPN in
methanolforthe fortificationofvegetable andricesampleswereprepared
vegetables was added 1 mL of a spiking solution. After setting aside for
24 h, the vegetable leaves were incubated in 5 mL of methanol for 10 min
with four vigorous shakes and then filtered through Whatman no. 1 filter
paper. The container and the residues were rinsed with 5 mL of methanol
and filtered, and the filtrate was combined with the previous filtrate.
due was resuspended in 10 or 100 mL of 20% methanol-PBS (200 mM).
The extract was analyzed by the indirect ELISA.
Analysis of Spiked Samples by GC. The spiked samples for GC
Daegu, Korea. The sample preparation procedure involved homogeniza-
on Florisil SPE with acetone/hexane (1:4). Quantification of EPN was
carried outbyGCusingaDB-5column(30m ? 0.25mmi.d., 0.25μm df)
and ECD as the detector. The injection port temperature was 260 ?C, and
the detectortemperature was 280?C.Oventemperature was programmed
from80?C (heldfor 2 min) to280 ?C ata rate of10 ?C/min. Nitrogengas
at 1.0 mL/min was used as the carrier gas.
RESULTS AND DISCUSSION
Production of MAbs. Among the six antisera from the mice
previously injected with hapten A-KLH or hapten B-KLH,
antisera M-1 (from hapten A-KLH) and M-5 (from hapten
B-KLH) with the highest pAb titer (1.313 at 1/50000 dilution
of the two antisera selected, a single well was selected, which
screening of the hybridoma clones by indirect ELISA resulted in
theselectionofthreehybridomas, 4Aand 6AfromhaptenA and
8B from hapten B.
Competitive Indirect ELISA. The results of the experiments
carried out to select the most suitable immunoreagents and their
appropriate concentrations using various combinations of anti-
body dilutions and various amounts of the coating antigens are
presented in Table 1. Only results with relatively high perfor-
mance were presented. Antibody 8B diluted 1/1000 and the
coating antigen hapten E-OVA at 500 ng/well were selected as
the most suitable on the basis of the IC50value that was the
lowest. Hapten heterology is commonly used to eliminate pro-
blems associated with the strong affinity of the antibodies to the
spacer arm that leads to no or poor inhibition by the target
compound (13,14). Of the five coating haptens prepared, hapten
B is homologoustothe immunizinghaptens in both positionand
structure of the bridge group, and haptens A, C, and D are
homologous to the immunizing haptens in position of the bridge
group, but heterologous in the structure of the bridge group.
Hapten E is the most heterologous because it is heterologous in
both position and structure of the bridge group. Therefore, the
result is in agreement with the strategy of using the competitor
with high hapten heterology to improve the assay sensitivity.
Because organic solvents are commonly used for extraction in
the analysis of pesticide residues in food and environmental
samples and pesticides are hardly soluble in aqueous solvent, it
Then, it is necessary to assess the effect of organic solvents on
ELISA performance at the competition step. The effects of
solvents (acetone, acetonitrile, methanol) on the ELISA system
were evaluated by preparing standard curves in buffers contain-
ing various amounts oforganic solvent.The results are presented
in Table 2. These solvents significantly influenced assay perfor-
methanol. It is interesting to note that although the maximum
absorbance was decreased to a large extent by increasing the
concentration of acetone and acetonitrile, the change was much
less with methanol. IC50values in the presence of acetone and
acetonitrile were much higher than those in the presence of
Combinations of Antibody and Coating Antigens
antibodydilutioncoating antigen (ng/well)
hapten B (100)
hapten C (200)
hapten E (1000)
hapten B (200)
hapten C (100)
hapten E (1000)
hapten B (200)
hapten C (200)
hapten E (500)
aMaximal absorbance (A), slope (B), IC50(C, ng/mL), and minimal absorbance
duplicates.bGoat anti-mouse IgG-HRP diluted 1/6000 was used.
Table 2. Influence of Organic Cosolvent, Phosphate Buffer, and pH of the
Assay Solution on Indirect Competitive ELISAa
aAssay conditions: antibody to hapten B-KLH, diluted 1/1000 with 10 mM
PBST; coating antigen, hapten E-OVA, 500 ng/well; goat anti-mouse IgG-HRP
diluted 1/6000. Maximal absorbance (A), slope (B), IC50(C, ng/mL), and minimal
absorbance (D) are values from the four-parameter sigmoidal fitting. The data are
the means of triplicates.bTime for color development.cData fitting was impossible
due to poor color development.dFinal concentration of phosphate ions of the
competition buffer containing 138 mM NaCl and 2.7 mM KCl.
5244J. Agric. Food Chem., Vol. 58, No. 9, 2010 Shim et al.
methanol. This trend was observed repeatedly in our previous
studies (7, 10, 11) and in other investigators’ studies (15-18).
Table 2 shows the lowest IC50value at 20% methanol.
Among the various assay parameters optimized, only the kind
of organic solvent has profound influence on assay performance.
Due to the nonpolar nature of EPN, it seems reasonable to
assume that hydrophobic interactions are important in the
antibody-analyte as well as antibody-coating antigen interac-
speculated that the ratio of antibody-analyte interaction to
antibody-coating antigen interaction is highest in the medium
containing methanol. Because increasing phosphate ion concen-
tration does not enhance the assay sensitivity (see below), the
higher interaction ratio appears to result from a certain factor
other than the polarity of the medium.
Table 2 also presents the effect of the phosphate ion (buffer)
concentration at the competition step on ELISA characteristics.
assay sensitivity, in contrast to the large effect observed in pAb-
based EPN ELISA (7). The optimum concentration selected was
200 mM phosphate, showing the lowest IC50value. Table 2 also
presents the effect of pH of assay solution on ELISA. The
physiological pH, pH 7.4, was selected as the optimum for the
Figure 2 shows a typical inhibition curve obtained under these
optimized conditions. The IC50value of the assay was 2.9 ng/mL
with a detection limit of 0.3 ng/mL (10% inhibition).
Competitive Direct ELISA. Antibody and enzyme tracer for
direct ELISA were optimized by testing various combinations of
enzyme tracers with haptens B-E were excluded in the direct
ELISA because they showed a very poor affinity to the antibody
optimum combination selected was antibody 6A from hapten A
diluted 1/7500 and the tracer hapten A-HRP prepared at a 50:1
hapten/protein molar ratio and diluted 1/2500. Hapten A is
homologous to immunizing hapten in both the position and
structure of the bridge group. We repeatedly observedthatdirect
ELISAs with homologous combinations showed better perfor-
mance compared to those with heterologous ones (7, 19, 20).
Here, we propose a hypothesis that there is no relationship
between the assay sensitivity and hapten heterology in direct
The effects of various parameters on the ELISA system are
presented in Table 3. Methanol was the most suitable cosolvent,
in agreement with the results of several other studies (21-23). It
may bespeculatedthattheratio ofantibody-analyte interaction
to antibody-enzyme tracer interaction is highest in the medium
containing methanol. Table 3 shows that assay sensitivity con-
tinues to improve with increasing concentration of methanol.
Twenty percent was selected as the optimum concentration,
because color development was so slow with 40% methanol.
The optimumconcentrationofthe phosphatebufferselectedwas
50 mM, which showed the strongest dynamic response (slope).
The physiological pH, pH 7.4, was selectedasthe bestone onthe
basis of the same criterion as for phosphate buffer.
Figure 3 shows a typical inhibition curve obtained under
optimized conditions. The IC50value of the assay was 0.6 ng/mL
with a detection limit of 0.09 ng/mL (10% inhibition).
Cross-Reactivity Studies. Several organophosphorus pesticides
were tested for cross-reactivity. Table 4 shows the cross-reactivi-
ties that were determined by both the indirect and direct assays,
Table 3. Influence of Organic Cosolvent, Phosphate Buffer, and pH of the
Assay Solution on Direct Competitive ELISAa
aAssay conditions: precoating with anti-mouse IgG (0.5 μg/well); antibody to
hapten A-KLH, diluted 1/7500 with 10 mM PBS; enzyme tracer, hapten A-HRP
(prepared at molar ratio 50:1, diluted 1:2500). Maximal absorbance (A), slope (B),
IC50(C, ng/mL), and minimal absorbance (D) are values from the four-parameter
cDatafitting was impossibledue to poorcolor development.dFinal concentration of
phosphate ions of the competition buffer containing 138 mM NaCl and 2.7 mM KCl.
Figure 2. ELISA standard curve for EPN by indirect competitive assay.
Assay conditions: antibody 8B raised against hapten B-KLH, diluted
1/1000; coating antigen, hapten E-OVA, 500 ng/well; goat anti-mouse
IgG-HRP, 1/6000; organic cosolvent, 20% methanol; phosphate buffer,
200 mM; pH of assay solution, 7.4. Each point of the standard curve
represents the mean of 16 determinations. Vertical bars indicate (
standard deviation about the mean.
ArticleJ. Agric. Food Chem., Vol. 58, No. 9, 20105245
expressed in a percentage of the IC50of EPN. The interference
with all of the pesticides tested was negligible.
been over three decades since immunoassays began to be devel-
oped for detecting pesticides, but studies in this field are still
actively carried out (24-27). It is now certain that most widely
used pesticides can be detected by immunoassay methods. How-
ever, contrary to the initial expectations, few pesticide immu-
noassays developed so far have been approved for official use.
Therefore, we have to draw attention to the limitations of
immunoassay methods and find ways to overcome the limita-
tions. The difficulties in resolution of matrix interference and in
standardization of reagents and procedures are among the
various limitations of pesticide immunoassays. With regard to
standardization of reagents, comparison of the capacity of pAb
and mAb in immunoassay is desirable.
be produced in unlimited quantities. We have developed several
pesticide ELISAs based on both pAbs and mAbs (9-12), and
comparison of their standard curves caused us to note that they
look different in some aspects. Thus, we began to presume that
of antibodies. We assumed that the additional merits of using
mAb may be better assay performance with regard to assay
specificity (28) and dynamic response. We attempted to test this
hypothesis in this study by comparing the assay performance of
mAb-based EPN ELISAs developed in this study with those of
the comparison to the ELISAs of organophosphorus pesticides
isofenphos and fenitrothion that we had developed previously
using both pAb and mAb (9-12).
Specificity in assay performances can be evaluated by examin-
ing the cross-reactivity of antibodies. The cross-reactivity values
of the two most cross-reactive pesticides in each of the pAb- and
mAb-based ELISAs for the three pesticides are presented in
Table 5. In the case of EPN, mAb-based ELISA in both indirect
and direct formats showed much better assay specificity than the
pAb-based ELISAs. In the case of isofenphos, the difference was
slight. The exception was the indirect ELISA for fenitrothion.
fenitrothion were slightly different in structure. Therefore, com-
somewhatinappropriatefortestingthe hypothesis. Morereliable
data would be those in McAdam et al.’s paper on the pAb- and
mAb-based ELISAs for fenitrothion (29). The corresponding
direct ELISA, respectively. Their result’s agreement with the
hypothesis is only with the direct ELISA. Their result’s disagree-
ment is higher if pesticide metabolites are included. Overall,
despite the clear correlation in the case of EPN ELISAs, we
should note that the data included in this study are not sufficient
to test the hypothesis that a mAb-based ELISA performs better
than a pAb-based ELISA with respect to assay specificity.
Theoretically, a mAb selected by clone isolation would not
necessarily show an exclusive affinity to the antigen. A clone
may be selected which secretes a mAb with a high affinity to an
epitope that is not unique to the antigen as well as to an epitope
that is unique to the antigen. Such a mAb would be less specific
than a mAb with an exclusive affinity only to an epitope that is
unique to the antigen. In the case of EPN, a mAb with a high
affinity only to the aromatic rings that is unique to EPN would
show better specificity than a mAb with a strong affinity to both
the aromatic rings and the thiophosphate group [P(dS);O-],
which is common among phosphorothioate organophosphorus
pesticides. Therefore, a mAb would show high specificity only
when an appropriate clone is selected.
For an analytical method to be sensitive, the detectable
concentrations must be low and dynamic response must be
strong. The sensitivity of an ELISA is usually expressed as the
IC50value, which is the median value of the detectable concen-
tration range. The dynamic response, also called “calibration
analyte concentration and, thus, is the slope of the calibration
curve (30). Higher dynamic response means narrower detectable
concentrations; however, it cannot be a problem because detect-
able concentrations can be easily adjusted by changing the
dilution factor of assay mixtures. As indicated by McAdam
et al. (29), the steepness of a standard curve is important for
sensitive detection by ELISAs; however, it is rarely presented in
the papers on ELISA development. We repeatedly observed that
of pAb-based ELISAs (9-12), which indicates higher dynamic
response with mAbs. Table 5 presents the slope of the standard
curves of mAb- and pAb-based ELISAs of the three pesticides.
than the corresponding polyclonal counterparts with no excep-
tion. Therefore, it may beconcluded that a mAb-based ELISA is
better than a pAb-based ELISA in dynamic response. The same
trend was observed in McAdam et al.’s paper (29).
For a wide range of concentrations to be detected, plateaus
must be clearly formed in the standard curves. Therefore, we
believe that the concept of dynamic response must include
clarity in the plateau formation, and an index indicating the
dynamic response in this aspect must be devised. We have
observed that the plateau is more clearly formed in the standard
curves of mAb-based ELISAs compared to pAb-based ELISAs.
Figure 3. ELISA standard curve for EPN by direct competitive assay.
Assay conditions: precoating agent, goat anti-mouse IgG, 0.5 μg/well;
antibody 6A raised against hapten A-KLH, diluted 1/7500; enzyme tracer,
cosolvent, 20% methanol; phosphate buffer, 50 mM; pH of assay solution,
7.4. Each point of the standard curve represents the mean of 16