Fiber Optic Biosensor for Cyclodiene Insecticides

Article abstract of DOI:10.1021/jf9701190

Chlorendic caproic acid (CCA) was used to synthesize hexachlorocyclopentadienefluorescein (FL) and bovine serum albumin (BSA) conjugates. Anti-CCA antibodies (CCA-Abs), which were raised against BSA-CCA and immobilized on quartz fibers, bound FL-CCA selectively and reversibly. Fluorescence generated by evanescent excitation of the bound FL-CCA was used to monitor the binding event. The affinity of CCA-Abs for FL-CCA (K_D = 1-9 nM) was calculated from the time courses of association and dissociation of FL-CCA. The cyclodiene insecticides chlordane, heptachlor, dieldrin, endrin, aldrin, and endosulfan competed with FL-CCA for binding to CCA-Abs and reduced fluorescence in a concentration-dependent manner with the following rank order: chlordane > heptachlor > dieldrin > aldrin > endosulfan. This fiber optic fluoroimmunosensor detects cyclodiene insecticides at the ppb level, has low cross-reactivity with γ-hexachlorocyclohexane, and does not detect (p,p'-dichlorodiphenyl)trichloroethane (DDT).

Full text of DOI:10.1021/jf9701190

3292
J. Agric. Food Chem. 1997, 45, 3292−3298
F iber Op tic Biosen sor for Cyclod ien e In secticid es
†
,†
‡
Kathleen E. Brummel, J eremy Wright,* and Mohyee E. Eldefrawi
Department of Biomedicinal Chemistry, School of Pharmacy, University of Maryland at Baltimore, 20
North Pine Street, Room 500, Baltimore, Maryland 21201, and Department of Pharmacology and
Experimental Therapeutics, School of Medicine, University of Maryland at Baltimore, 655 West Baltimore
Street, BRB 4-029, Baltimore, Maryland 21201
Chlorendic caproic acid (CCA) was used to synthesize hexachlorocyclopentadienefluorescein (FL)
and bovine serum albumin (BSA) conjugates. Anti-CCA antibodies (CCA-Abs), which were raised
against BSA-CCA and immobilized on quartz fibers, bound FL-CCA selectively and reversibly.
Fluorescence generated by evanescent excitation of the bound FL-CCA was used to monitor the
binding event. The affinity of CCA-Abs for FL-CCA (KD ) 1.9 nM) was calculated from the time
courses of association and dissociation of FL-CCA. The cyclodiene insecticides chlordane, heptachlor,
dieldrin, endrin, aldrin, and endosulfan competed with FL-CCA for binding to CCA-Abs and
reduced fluorescence in a concentration-dependent manner with the following rank order: chlordane
>
heptachlor > dieldrin > aldrin > endosulfan. This fiber optic fluoroimmunosensor detects
cyclodiene insecticides at the ppb level, has low cross-reactivity with γ-hexachlorocyclohexane, and
does not detect (p,p’-dichlorodiphenyl)trichloroethane (DDT).
Keyw or d s: Cyclodiene insecticides; fiber optic biosensor; fluoroimmunoassay; detection
INTRODUCTION
in soil from applications made 20 years ago (ATSDR,
1
1
989; Kilburn and Thornton, 1995; Rimkus and Wolf,
987; Van Wijnen and Stijkel, 1988; Murphy, 1995).
A continuing problem is that leaching into the envi-
The cyclodiene insecticides dieldrin, aldrin, hep-
tachlor, chlordane, endrin, and endosulfan are polychlo-
rinated hydrocarbon compounds. A characteristic poly-
chlorinated bicyclic ring system is found in each member
of this series. They were first synthesized via a Diels-
Alder reaction in the 1940s as an insecticide alternative
to DDT (Ungnade and McBee, 1958; ATSDR, 1989).
Cyclodiene insecticides were heavily used in agriculture
and were popular because of their broad spectrum
action, high insecticidal activity, and persistence at the
site of exposure. In addition, they were highly effective
against certain pests such as fire ants and termites
ronment occurs from hazardous waste sites which
contain unknown amounts of cyclodiene insecticide
residues that are unevenly distributed. Due to the over
abundance of chemical waste containing mixtures of
many pesticides, a quick and highly specific detection
method is warranted in order to locate and properly
dispose of these toxic compounds. A recent special
report discusses the current status and future applica-
tions for monitoring pesticide residues in the environ-
ment (Kr a¨ mer, 1996). Several analytical methods are
available for analysis of cyclodiene insecticides. Extrac-
tion techniques along with gas chromatography/mass
spectrometry (GC/MS) have been employed on these
compounds (ATSDR, 1989; Rimkus and Wolf, 1987;
Sovocool et al., 1977; Van Wijnen and Stijkel, 1988).
Because γ-aminobutyric acid (GABA) receptor is the
molecular target for cyclodienes, a ligand binding assay
to this receptor was used to measure the amount of
R-endosulfan in blood samples (Saleh et al., 1993).
Enzyme-linked immunosorbant assay (ELISA) has been
developed for the detection of organochlorine pesticide
residues; the limit of detection was 0.01-0.03 ppm
(Wigfield and Grant, 1992). A radioimmunoassay for
dieldrin and aldrin was developed, using their carboxyl-
substituted analogues bound to horse serum albumin
(HSA) to produce polyclonal antibodies (Abs) (Langone
and Van Vunakis, 1975). The Ab specificity was di-
rected toward the distal part of the hapten, i.e. the
hexachlorobicyclic ring system. Dieldrin was detected
at 150 pg, aldrin at 700 pg, and heptachlor and chlor-
dane at higher concentrations. Little cross-reactivity
was seen with other chlorinated hydrocarbons such as
lindane. A cyclodiene-sensitive monoclonal Ab (mAb),
derived from an aldrin analogue, was also produced and
used in pharmacological studies on GABA receptor
(Esser et al., 1991).
(EPA, 1974).
The high lipid solubility of polychlorinated hydrocar-
bon insecticides caused them to be readily taken up and
bioconcentrated in animal fat depots, such as adipose
tissues, nerves, and brain (Geyer et al., 1993). Hep-
tachlor was reported to bioaccumulate 200-37000 times
from water into hydrobiota (WHO, 1984). Aldrin, hep-
tachlor, and chlordane were shown to be highly resistant
to degradation in soil and persist for years. Endosulfan
was shown to be biodegradable and is still used today.
Heptachlor epoxide and dieldrin persisted in both
aerobic and anaerobic environments (Tu and Miles,
1
976; Pfaffenberger et al., 1994). The biomagnification
throughout the food chain and persistence in the
environment (Geyer et al., 1993; Metcalf et al., 1973;
Cole et al., 1976) rendered cyclodiene insecticides eco-
logically harmful and environmentally hazardous (Wurst-
er, 1971; Rimkus and Wolf, 1987; Van Wijnen and
Stijkel, 1988; Bloomquist, 1992; Kilburn and Thornton,
1
995). This led to a ban on the use of cyclodiene
insecticides in the 1970s by the Environmental Protec-
tion Agency (EPA). However, residues can still be found
*
Author to whom correspondence should be ad-
dressed [telephone (410) 706-2114; fax (410) 706-0346;
e-mail wright@pharmacy.ab.umd.edu].
†
School of Pharmacy.
School of Medicine.
‡
Fiber optic immunoassays have been shown to be
S0021-8561(97)00119-2 CCC: $14.00
© 1997 American Chemical Society

Biosensor for Cyclodiene Insecticides
J. Agric. Food Chem., Vol. 45, No. 8, 1997 3293
F igu r e 1. Synthetic scheme of CCA-hapten (2) and CCA-BSA antigen (3). Fluorescent probe, FL-CCA (4), is synthesized via
chlorendic anhydride (1) with FED.
highly specific and sensitive toward target analyte(s)
and are promising for future field use (Rogers et al.,
measure the fluorescence generated by the fluorescent probe
FL-CCA), when bound to the Abs on the surface of the quartz
(
fiber within the evanescent zone, which extends 1000 Å from
the surface. The quartz fiber (1 × 60 mm) (Wale Apparatus,
Hellertown, PA) was placed inside a flow cell of 46 µL capacity,
which was exchanged by selected solutions every 14 s. The
instrument used a 485 nm excitation filter with a 20 nm band
pass and a 530 nm emission filter with a 30 nm band pass. A
photon source was generated with an InGaN/AlGaN/GaN blue-
light-emitting diode operated at 20 mA (Nichia Chemical
Industries LTD, J apan) and was detected with a Hamamatsu
S-1087 silicon detector.
1
991; Anis and Eldefrawi, 1993; Eldefrawi et al., 1993).
The fiber optic biosensor consists of Abs immobilized
on the surface of a quartz fiber and is used to detect
fluorescent analytes by means of the evanescent wave,
which causes excitation of the bound fluorescent ana-
lyte. Measurement of the target analyte(s) is achieved
by competitive displacement of the fluorescent analyte
by selective transduction of a parameter of the biomol-
ecule-analyte reaction into a quantifiable signal (By-
field and Abuknesha, 1994).
In this study, a fiber optic evanescent fluorosensor
was developed using, as biological sensing elements,
polyclonal Abs isolated from rabbit immune serum and
immobilized covalently onto the quartz fibers. A fluo-
rescent conjugate of the target compound (designated
as the probe), which fluoresces when bound to the Ab,
was used as the optical signal generator. The presence
of a cyclodiene insecticide in the flow buffer competi-
tively displaced the fluorescent probe, thereby reducing
the fluorescence. This biologically optimized molecular
recognition (Ab-Ag interaction) was used to detect and
quantitate cyclodiene and other chlorinated hydrocar-
bons.
Ch em ica l Str u ctu r e Elu cid a tion . Nuclear magnetic
resonance (NMR) spectra were obtained by a General Electric
QE-300 Fourier transform spectrometer. Thin layer chroma-
tography (TLC) was run with analytical silica gel plates
(Fisher Scientific, Pittsburgh, PA). Elemental analysis (C, H,
N, Cl) was performed on the synthesized compounds at
Atlantic Microlab, Inc., Norcross, GA.
Syn t h esis of 6-(1,4,5,6,7,7-H exa ch lor ob icyclo[2.2.1]-
h ep ta n e-2,3b-su ccin im id o)ca p r oic Acid [Ch lor en d ic Ca -
p r oic Acid (CCA, 2), F igu r e 1).] 6-Aminocaproic acid (0.7
g, 5.3 mmol) was added to a solution of chlorendic anhydride
(1) (2 g, 5.4 mmol) and equimolar triethylamine (TEA) in 15
mL of toluene in a round bottom flask fitted with a Dean-Stark
water separator. The solution was refluxed for 8 h, toluene
was removed under vacuum, and the residue was triturated
with 20 mL of 1% HCl. The remaining tan solid was collected
by vacuum filtration and crystallized from a hexane/methylene
chloride solution (5:1). Yield was 2.3 g (88%) of white
MATERIALS AND METHODS
crystals: mp 150-151 °C; TLC (4/1 ethyl acetate/methanol/
Ch em ica ls. Chlordane, dieldrin, aldrin, heptachlor, endrin,
lindane, DDT, endosulfan, and chlordecone (99.3-100% puri-
ties) were obtained from ChemService (West Chester, PA).
Chlorendic anhydride (97% purity), 6-aminocaproic acid, and
ethylenediamine dihydrochloride as well as other chemicals
were from Aldrich Chemical (Milwaukee, WI). Fluorescein
isothiocyanate (90% purity), bovine serum albumin (96-98%
purity), casein, goat anti-rabbit IgG (whole molecule), alkaline
phosphatase conjugate, p-nitrophenyl phosphate, and other
chemicals used in the ELISA were purchased from Sigma and
Sigma Immuno Chemical Co. (St. Louis, MO). Commercial
anti-heptachlor neat serum (from BioDesign International,
Kennebunk, ME) was purified with a Protein A affinity column
from Pierce (Rockford, IL).
1
1
1
1
% acetic acid) R
f
)0.7; H NMR (DMSO-d
6
): δ 1.2 (m, 2H),
.3 (m, 2H), 1.4 (m, 2H), 2.1 (t, 2H), 3.28 (t, 2H), 4.06 (s, 2H),
2.02 (s, 1H). Anal. C, H, N, Cl.
Syn th esis of CCA-BSA (3) a n d F lu or escein (4) Con -
ju gates (Figu r e 1). Fluoresceinthiocarbamoylethylenediamine
(FED) was synthesized according to Pourfarzaneh et al. (1980).
Briefly, ethylenediamine dihydrochloride was reacted with
fluorescein isothiocyanate (5:1 molar ratio) in methanol with
a small amount of TEA. An orange precipitate (4) was
collected, washed with methanol (70% yield), and reacted
directly with an equimolar amount of chlorendic anhydride
(1) in dimethylformamide (DMF) and TEA. The reaction
mixture was acidified with 0.5 M HCl, and the orange
precipitate was isolated by filtration. TLC (5/1 ethyl acetate/
Ap p a r a tu s. Experiments were performed on the ORD
portable fluorometer (ORD, Inc., North Salem, NH), as de-
scribed by Rogers et al. (1989). The instrument was used to
methanol/1% acetic acid) showed an R
) 0.8. NMR showed
f
all characteristic peaks of hexachlorocyclopentadiene fluores-
cein (FL-CCA, compound 4). Carboxylic hapten (2) (0.96 g,

3294 J. Agric. Food Chem., Vol. 45, No. 8, 1997
Brummel et al.
0
.2 mmol), N-hydroxysuccinimide (NHS) (0.2 mmol), and
-
4
dicyclohexylcarbodiimide (DCC) (0.45 g, 2.2 × 10 mol) were
dissolved in 1.0 mL of anhydrous DMF. The solution was
stirred for 3.5 h at 15 °C, and the precipitated dicyclohexylurea
was removed by centrifugation. The DMF supernatant was
added to a solution of 200 mg of BSA, 20 mL of water, and 4.2
mL of DMF. The white reaction mixture was stirred gently
at 4 °C for 24 h to complete the conjugation. Extensive dialysis
against four changes of 4 L of 10 mM phosphate-buffered
solution (PBS) over 4 days was used to purify the conjugate.
The product was lyophilized, and the resulting white solid was
analyzed. C, H, N, Cl analysis showed approximately 37
molecules of CCA coupled to 1 molecule of BSA.
In d ir ect ELISA An a lysis. Microtiter plates (96-well) were
coated with 50 µL of 5 µg/mL CCA-BSA in a 0.1 M carbonate
buffer and stored at 4 °C overnight (18 h). The plates were
then washed twice with Tris-buffered saline (TBS) and 0.5%
Tween 20 (wash buffer), and then all the wells were blocked
with 300 µL/well of 10% skim milk with 2% sodium azide
(blocking buffer) for 2 h. Subsequently, the plates were
washed four times with wash buffer and incubated with 50
µL/well of a 1:1000 dilution of commercial anti-heptachlor Ab
in blocking buffer for 1 h. After 1 h the plates were again
washed four times with wash buffer and treated with 50 µL/
well of goat anti-rabbit IgG (whole molecule) alkaline phos-
phatase conjugate in blocking buffer for 1 h. The plate was
rinsed six times with wash buffer and developed with p-
nitrophenyl phosphate in a 50 mM carbonate buffer and 5 mM
F igu r e 2. Specificity of the fluorescent signal transmitted by
quartz fibers coated with anti-CCA Abs when perfused with
flow buffer containing 10 nM FL-CCA (9, 0), 10 nM FL-R-
BGT (b), 10 nM FL-benzoylecgonine (2), or 10 nM FL-
atrazine ([). The quartz fibers coated with anti-heptachlor
Abs fluoresce when perfused with buffer containing 10 nM
MgCl
solution at 50 µL/well. A color change was detected
2
immediately, and the OD values at 405 nm and 630 nm were
determined on an EL312e Biokinetics plate reader. CCA-
BSA (3) was detected by commercial anti-heptachlor at dilu-
tions of 1:8000, an indication of successful formation of a CCA-
BSA conjugate. The same procedure was used to synthesize
CCA-casein for use in ELISA assays.
a
FL-CCA ( ).
4
°C for 14 h. The fibers were then washed with 50 mM PBS
and stored until use in PBS containing 0.05% of NaN
3
.
F lu or escen ce Mea su r em en ts. The Ab-coated fibers were
mounted in the flow cell of the ORD fluorometer and perfused
with PBS containing 0.1% casein (100 µg/mL) for a period of
P r ep a r a tion of P olyclon a l Ab to CCA. A New Zealand
White rabbit (Hazelton, PA) was injected intramuscularly at
three or four sites with 1 mL containing 1:1 Freund’s complete
adjuvant and 1 mg/mL CCA-BSA conjugate (3). The rabbit
was boostered four times intramuscularly with 1 mg/mL CCA-
BSA and Freund’s incomplete adjuvant in a 1:1 ratio (0.8-1
mL per booster) over a period of 3 months. Immune serum
5
min to establish baseline fluorescence. FL-CCA (10 nM)
was added to the buffer, and the fluorescence generated in the
evanescent zone, as a result of binding of CCA-FL to the Ab
on the surface of the fiber, was transmitted through the fiber
to the silicon detector, translated into voltage, and recorded.
(20 mL) was collected through cardiac puncture, and Ab titers
were determined by ELISA. A CCA-casein conjugate and
BSA (all at 5 µg/mL) were coated on 96-well microtiter plates
and tested with immunized rabbit serum and control rabbit
serum. Titers were detected with goat anti-rabbit IgG (whole
molecule) alkaline phosphatase conjugate and p-nitrophenyl
phosphate as a substrate. Strong responses were obtained at
RESULTS
Sp ecificity of th e F lu or oim m u n osen sor . The
binding of FL-CCA to anti-CCA Ab-coated fibers re-
sulted in a fluorescent signal, the amplitude of which
increased with higher FL-CCA concentrations. At 10
nM FL-CCA the fluorescence increased with higher
anti-CCA Ab density on the fiber. When the anti-CCA
Ab-coated fibers were perfused with flow buffers, each
containing 10 nM of FL-CCA, FL-atrazine, FL-R-
bungarotoxin (R-BGT) or FL-benzoylecgonine, strong
fluorescence was generated only with FL-CCA (Figure
1
:64000 for the CCA-casein-coated wells and less response
to the BSA. No response was seen for the unimmunized serum
with CCA-casein or BSA.
P u r ifica tion of An ti-CCA Abs. The rabbit antiserum was
purified using an ImmunoPure Plus IgG purification kit
following the manufacturer protocol. Briefly, the column was
first equilibrated with 5 mL of ImmunoPure binding buffer,
and then 4 mL of serum, diluted 1:1 with binding buffer, was
added. The column was immediately washed with 15 mL of
binding buffer in order to remove any non-IgG proteins. The
bound IgG was then eluted with 5 mL of elution buffer, and 5
2
). Very low fluorescence (<10%) was generated by the
other FL reagents, most likely due to nonspecific bind-
ing. Fibers coated with the commercially available anti-
heptachlor Abs had lower binding capacity of FL-CCA
×
1 mL fractions were collected. Absorbance was monitored
(Figure 2) and slightly higher nonspecific binding of the
at 280 nm in a Beckman DU-50 spectrophotometer. Two
concentrated IgG fractions were collected. The purified IgG
was dialyzed against 3 × 4 L of 10 mM PBS over a period of
other three fluorescent reagents (data not shown).
Nonspecific binding to the anti-heptachlor Ab-coated
fiber amounted to 28% of FL-CCA binding compared
to only 10% for the anti-CCA Ab-coated fiber. These
values were used to correct for specific binding in
subsequent experiments. Fibers coated with control
rabbit IgG or BSA failed to bind FL-CCA and trans-
mitted no fluorescence above the base line level. Per-
fusion with 10 nM FL-CCA rapidly increased the
fluorescence transmitted by anti-CCA Ab-coated fibers
and reached a steady state in 30-40 min, while removal
of FL-CCA from the flow buffer decreased fluorescence
in a time-dependent manner (Figure 3). Since fluores-
2
days. Subsequent analysis for protein concentration (Lowry
et al., 1951) indicated 12.48 mg/mL.
Cova len t Im m obiliza tion of Abs on th e Qu a r tz F iber s.
This was accomplished by a modified protocol of Bhatia et al.
(
1989). The fibers were washed successively in HNO
3 2 4
, H SO ,
and HCl-EtOH, rinsed in double-distilled water, and dried
at 70 °C. The clean fibers were then silanized with (3-
mercaptopropyl)methoxysilane to introduce an active SH
group, and then reacted with the heterobifunctional cross-
linking reagent N-[(γ-maleimidobutyryl)oxy]succinimide. The
fibers were then individually incubated in 40 µL of Ab solution
(
50 µg/mL) placed inside a microcapillary tube (Kimax 51) at

Biosensor for Cyclodiene Insecticides
J. Agric. Food Chem., Vol. 45, No. 8, 1997 3295
competitive inhibition data (Table 1), using the Cheng-
Prusoff (1973) equation [Ki ) IC50/(1 + [L]/KD)], where
[L] is the concentration of the fluorescent probe. To
calculate the Ki values, [L] was 10 nM and KD was 1.9
nM for anti-CCA. Chlordane and heptachlor had Ki
values of 1.38 and 1.66 nM, respectively; followed by
dieldrin with a Ki of 7.59 nM and aldrin with a Ki of
1
8 361 nM.
Cr oss-Rea ctivity w ith Oth er Ch lor in a ted Hy-
d r oca r bon s. Because the cyclodiene insecticides mimic
CCA, this compound was used to generate the antisera
and the fluorescent probe. It was thought that the anti-
CCA Ab would have highest affinity for CCA, but would
cross-react with structurally related chlorinated hydro-
carbon compounds. The concentration selected to gen-
erate cross-reactivity data was 0.1 µM of CCA, which
gave complete inhibition of fluorescence, but significant
partial inhibition by aldrin, the least inhibitory com-
pound. The anti-CCA Ab immunosensor showed 60 and
6
4% cross-reactivity to heptachlor and chlordane, re-
spectively, but only 20 and 50% to endrin and dieldrin,
respectively. The cross-reactivity at different concen-
trations of chlordane levels changes the ranking of
chlordane and heptachlor. This change still results in
detection of heptachlor and chlordane better than di-
eldrin and aldrin. Its cross-reactivity to endosulfan was
very low (9%). DDT, lindane, and chlordecone showed
cross- reactivities ranging from 0 to 8% (Table 1). Cross-
reactivities of immunosensors using the commercially
available anti-heptachlor Abs for the cyclodiene insec-
ticides were different (Table 1), with higher cross-
reactivity for the cyclodiene insecticides than for CCA.
F igu r e 3. Time course of fluorescent increase as a result of
association of FL-CCA (10 nM) with anti-CCA-coated quartz
fibers (50 µg/mL). After 20 min, FL-CCA was withdrawn
from the flow buffer and the dissociation time course monitored
for 10 min.
cence originates in the evanescent zone, it is safe to
assume that the increase and decrease in fluorescence
represent binding to, and dissociation of FL-CCA from,
the fiber. All studies were performed with single-use
fibers.
Kin etics of Associa tion a n d Dissocia tion of F L-
CCA w ith F iber s Coa ted w ith An ti-CCA Abs. The
apparent association rate constant (kapp) was calculated
from the slope of the initial 15 min of the rising portion
of the linear plot of ln(Flss/Flss - FLt) vs time. Flss is
the steady state value measured after 60 min; t is the
time interval at which fluorescence was measured. The
dissociation rate constant (k-1) was calculated by de-
termining the halftime (T1/2) of dissociation following
removal of FL-CCA from the flow buffer after reaching
the steady state. The semilogarithmic plot of fluores-
cence versus time was then used to obtain T1/2 followed
by k-1, calculated from the formula k-1 ) 0.693/T1/2. The
association rate constant k1 was calculated from the
formula k1 ) (kapp - k-1)/[FL-CCA]. The affinity
constant KD was calculated using the formula KD ) k-1/
k1. The KD value for FL-CCA binding to anti-CCA Abs
was determined to be 1.9 nM. For the commercial anti-
heptachlor Abs we saw two dissociation components.
The major component was the slow dissociating one with
T1/2 ) 13 min and KD) 2.8 nM.
DISCUSSION
The specificity of the fiber optic biosensor coated with
anti-CCA Abs is evidenced by the strength of the
fluorescent signal produced when it binds FL-CCA. The
unrelated chemicals FL-R-BGT, FL-benzoylecgonine,
or FL-atrazine (Figure 2) produce low, if any, fluores-
cence at the same concentrations. The signal is time-
dependent and is detectable in 1 min. Binding of the
FL-CCA to the CCA-Ab-coated fiber is reversible
(Figure 3), as expected of the Ag-Ab interaction. The
anti-CCA Ab detects chlordane by inhibition of the
fluorescence signal in a dose dependent manner (Figure
4A). As little as 0.1 nM chlordane is easily detectable
and within 1 min.
It is evident that this fluoroimmunosensor can detect
and quantitate cyclodiene insecticides (Figures 4 and 5
and Table 1). It utilizes the same competitive immune
assay strategy described previously for detection of
polychlorinated biphenyls (Zhao et al., 1995) and
imazethapyr herbicide (Anis and Eldefrawi, 1993). It
is highly sensitive, detecting 0.3 ppb FL-CCA (equiva-
lent to 1 nM) and chlordane and heptachlor at 1 nM
(Figure 5). Dieldrin and aldrin are detected at 10 and
1000 times higher concentration; therefore the biosensor
has lower sensitivity toward these compounds. The
sensitivity of the biosensor is comparable to that of a
radioimmunoassay, reported to detect 0.15 ppb dieldrin
and 0.7 ppb aldrin (Langone and Van Vunakis, 1975).
Their detection levels of heptachlor and chlordane by
the radioimmunoassay were 13 and 26 times less than
dieldrin. The hapten used to synthesize the Ag to raise
Abs used in that radioimmunoassay was a carboxylic
acid derivative of dieldrin. An ether derivative of aldrin
was also used to raise mAbs that bound chlordane and
endosulfan (Esser et al., 1991). This mAb cross-reacted
Com p etitive Disp la cem en t of F L-CCA by Cy-
clod ien e In secticid es. The affinity of the immun-
osensor for cyclodiene insecticides was determined from
a plot of fluorescence vs time. Fluorescence resulting
from binding of FL-CCA to anti-CCA Ab-coated fibers
was reduced in the presence of chlordane (0.1-1000 nM)
in a concentration dependent manner (Figure 4A).
Linear transformation of the percent inhibition of bind-
ing at steady state allowed accurate determination of
the concentration required to obtain 50% inhibition (i.e.
IC50) (Figure 4B). The following was the rank order of
competitive inhibition of FL-CCA binding by the in-
secticides: CCA > chlordane > heptachlor > dieldrin
>
aldrin (Figure 5). The IC50 values for the insecticides
was calculated from the linear transformation of the

3296 J. Agric. Food Chem., Vol. 45, No. 8, 1997
Brummel et al.
F igu r e 4. Concentration-dependent inhibition of FL-CCA binding to anti-CCA-coated quartz fibers by chlordane (A). The
logarithmic transformation of inhibition versus percent fluorescent signal (B).
Ta ble 1. Cr oss-Rea ctivities of An ti-CCA a n d Com m er cia l
An ti-Hep ta ch lor -Coa ted F iber s w ith In secticid es
cross-reactivitiesb
Ki (nM) of
anti-CCA Ab
anti-CCA
Abs
anti-heptachlor
Abs
a
compound
CCA
0.16 ( 0.013
1.38 ( 0.011
1.66 ( 0.004
7.59 ( 0.003
>1000
100
63.7
60.1
51.0
7.1
20.2
9.0
8.3
100
123
121
114
135
90
38
0
0
0
chlordane
heptachlor
dieldrin
aldrin
endrin
>1000
endosulfan
lindane
chlordecone
DDT
>1000
>1000
>1000
2.4
0
>1000
a
Mean Ki values of duplicate or triplicate experiments ( SD,
b
i.e. the mean value of all the sums of all measurements. Cross-
reactivity values of chlorinated insecticides (each obtained from
triplicate experiments) are calculated as follows: the rate of
fluorescence decrease caused by the addition of 100 nM of CCA to
the flow buffer is considered as 100. The initial rate of fluorescence
decrease in the presence of 100 nM of each compound is divided
by that for CCA to obtain the cross-reactivity index. A cross-
reactivity value above 100 indicates detection of the compound at
lower concentrations than CCA.
F igu r e 5. Dose-dependent competitive inhibition of fluores-
cence of anti-CCA-coated fibers (50 µg/mL) by CCA (0), aldrin
(
0), dieldrin (b), heptachlor (2), and chlordane ([). Each
The commercial anti-heptachlor Ab has lower affinity
for FL-CCA (Figure 2) but higher affinity for all the
cyclodiene insecticides (Table 1). The higher cross-
reactivity means that the anti-heptachlor Abs, although
having lower affinity for the probe, have higher sensi-
tivity in detecting cyclodiene insecticides. Thus, while
the use of CCA as a hapten to generate anti-cyclodiene
Abs is successful, other antigens may generate Abs with
even higher sensitivity for cyclodiene insecticides. The
cross-reactivity of the biosensor with lindane and chlo-
rdecone, although low (<10%), is still detectable. Since
cross-reactivity was tested at 0.1 µM concentrations (40
ppb) of the different chlorinated hydrocarbons, it is
possible that this biosensor can yield false positive
results if either of these insecticides is present at ppm
concentrations. This problem may be overcome by
increasing the threshold cut off level 10-fold (i.e. 400
ppb). There are obvious approaches to improve sensi-
tivity and selectivity of the biosensor and reduce the
possibility of false-positive results. One is to use mAbs
selected for their high affinities for each cyclodiene
symbol represents the mean of duplicate or triplicate experi-
ments with SD < 14%.
with heteroadamantanes, which are nonchlorinated and
bear little structural resemblance to cyclodiene insec-
ticides. However, they bind to the same site on the
GABA receptor that binds cyclodiene insecticides. Thus
the carboxylic acid-derived Abs appear to be selective.
Heptachlor and chlordane contain a hexachlorobicy-
clic ring system similar to that of CCA (Figure 6) and
also have a five-membered ring system with one or two
chlorine atoms. It is evident that heptachlor and chlo-
rdane compete with FL-CCA binding to the anti-CCA
Ab-coated fiber better than the other cyclodiene insec-
ticides tested (Table 1). However, the Ki values of
chlordane (Ki ) 1.38 nM) and heptachlor (Ki ) 1.66 nM)
are 10 times lower than that of CCA (Ki ) 0.16 nM).
Aldrin, dieldrin, and endrin, with more complex stere-
ochemical structure and less likelihood of being recog-
nized by the anti-CCA Ab, have even lower Ki values
(Table 1).

Biosensor for Cyclodiene Insecticides
J. Agric. Food Chem., Vol. 45, No. 8, 1997 3297
posable upon laboratory generated curves. Detection
of cyclodiene insecticides in soil is expected to yield a
generic method. Until highly selective mAbs are pro-
duced, samples giving positive values by the above
method would also have to be assayed by GC/MS or
HPLC if identification of the individual cyclodiene
insecticide is required.
ABBREVIATIONS USED
CCA, chlorendic caproic acid; FL-CCA, hexachloro-
cyclopentadiene fluorescein.
ACKNOWLEDGMENT
We thank Dr. Amira T. Eldefrawi and Ms. Vania I.
Cortes for assistance in preparation of the manuscript.
LITERATURE CITED
F igu r e 6. Structure of several cyclodiene insecticides.
Agency for Toxic Substance and Disease Registry U. S. Public
Health Service (ATSDR). In Toxicological Profile for Aldrin/
Dieldrin. Oak Ridge National Laboratory, Eds.; Dynamic
Corp., 1989.
Anis, N. A.; Eldefrawi, M. E. Reusable fiber optic immunosen-
sor for rapid detection of imazethapyr herbicide. J . Agric.
Food Chem. 1993, 41, 843-848.
Bhatia, S. K.; Shriver-Lake, L. C.; Prior, K. J .; Georger, J . H.;
Calvert, J . M.; Bredehorst, R.; Ligler, F. S. Use of thiol
terminal silanes and heterobifunctional crosslinkers for
immobilization of antibodies on silica surfaces. Anal. Bio-
chem. 1989, 178, 408-413.
insecticide. Higher affinity of the mAb for a cyclodiene
insecticide than other cyclodienes, which are close in
structure, requires selection of the ideal hapten. This
strategy to improve sensitivity and selectivity of the
biosensor are currently under study.
Advantages of this solid phase fluoroimmunosensor
over other immune assays include the use of less
hazardous materials compared to radioimmunoassay
(Langone and Van Vunakis, 1975) and speed compared
to both radioimmunoassay and ELISA (Engvall, 1980).
This particular fiber optic sensor is best suited for single
measurements. Regeneration of the sensor takes time
and eliminates the advantages of quick screening if a
large number of samples is involved. However, there
are other biosensors such as the light addressable
potentiometric sensor (LAPS) with high throughput that
was used successfully for the detection of organophos-
phorous insecticides (Fernando et al., 1993). As shown
in Figure 4A, chlordane reduces not only the steady-
state fluorescence, but also the rate of association, in a
concentration-dependent manner. The reduction in the
initial rate of fluorescence increase was shown to
parallel the reduction in steady-state fluorescence (Anis
and Eldefrawi, 1993; Devine et al., 1995). Thus, rate
measurements obtained in minutes could be used for
quick screening purposes. Such rate measurements of
cocaine alkaloids in coca leaf extracts gave values
comparable to those obtained by fluorescence polariza-
tion and gas chromatography (Toppozada et al., 1997).
The fiber optic fluorometer is a portable system, and
variants of the instrument are available commercially
from Research International, Woodinville, WA (Ligler
et al., 1993). Such portable instruments can be used in
on-site analysis for rapid screening of large numbers of
samples to define the boundaries of a contaminated site.
It is obvious that the immunosensor will identify cyclo-
diene insecticides as a group. However, biosensors
using mAbs raised against different haptens would have
various sensitivities for the different cyclodienes. The
advantages from previous studies on other environmen-
tal pollutants demonstrated a quick screening tech-
nique. Little sample pretreatment was necessary to
assay soil samples for the detection of imazethapyr (Anis
and Eldefrawi, 1993; Zhao et al., 1995). This was
performed by extracting the soil with water, sediment-
ing the particles, and then filtering and buffering the
extract with PBS. Dose-response curves for imaze-
thapyr obtained from these soil extracts were superim-
Bloomquist, J . R. Intrinsic lethality of chloride-channel-
directed insecticides and convulsants in mammals. Toxicol.
Lett. 1992, 60, 289-298.
Byfield, M. P.; Abuknesha, R. A. Biochemical aspects of
biosensors. Biosens. Bioelect. 994, 9, 373-400.
Cheng, Y.; Prusoff, W. H. Relationship between the inhibition
constant (K
i
) and the concentration of an inhibitor that
causes a 50% inhibition (IC50) of an enzyme reaction.
Biochem. Pharmacol. J . 1973, 22, 3099-4106.
Cole, L. K.; Metcalf, R.; Sanborn, J . R. Environmental fate of
insecticides in terrestrial model ecosystems. Int. J . Environ.
Studies 1976, 10, 7-14.
Devine, P. J .; Anis, N. A.; Wright, J .; Kim, S.; Eldefrawi, A.
T.; Eldefrawi, M. E. A fiber optic cocaine biosensor. Anal.
Biochem. 1995, 227, 216-224.
Eldefrawi, M. E.; Eldefrawi, A. T.; Anis, N. A.; Valdes, J . J .
Reusable fiber optic biosensors for detection of drugs and
toxicants. In Uses of immobilized Biological Compounds for
Detection, Medical, Food and Environmental Analysis; Guil-
bault, G. G., Mascini, M., Eds.; NATO ASI Ser. E; Kluwer
Academic: Dordrecht, The Netherlands, 1993, pp 387-395.
Engvall, E. Enzyme immunoassay ELISA and EMIT. Methods
Enzymol. 1980, 70, 419-439.
Environmental Protection Agency (EPA). Pesticide products
containing heptachlor or chlordane: intent to cancel regis-
trations. Fed. Register 1974, 39, 41298-41300.
Esser, T.; Karu, A. E.; Toia, R. F.; Casida, J . E. Recognition of
tetramethylenedisulfotetramine and related sulfamides by
the brain GABA-gated chloride channel and a cyclodiene-
sensitive monoclonal antibody. Chem. Res. Toxicol. 1991,
4
, 162-167.
Fernando, J . C.; Rogers, K. R.; Anis, N. A.; Valdes, J . J .;
Thompson, R. G.; Eldefrawi, A. T.; Eldefrawi, M. E. Rapid
detection of anticholinesterase insecticides by a reusable
light addressable potentiometric biosensor. J . Agric. Food
Chem. 1993, 41, 511-516.
Geyer, H. J .; Scheunert, I.; Rapp, K.; Gebefugi, I.; Steinberg,
C.; Kettrup, A. The relevance of fat content in toxicity of
lipophilic chemicals to terrestrial animals with special
reference to dieldrin and 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD). Ecotoxicol. Environ. Saf. 1993, 26, 45-60.

3298 J. Agric. Food Chem., Vol. 45, No. 8, 1997
Brummel et al.
Kilburn, K. H.; Thornton, J . C. Protracted neurotoxicity from
chlordane sprayed to kill termites. Environ. Health Perspect.
Saleh, M. A.; Zied, M. A.; El-Baroty, G.; Abdel-Reheim, E.;
Abdel-Rahman, F.; Wallace, C.; El-Sebae, A. H.; Blancato,
J . N. Gamma aminobutyric acid radioreceptor-assay a
possible biomarker for human exposure to certain agro-
chemicals. J . Environ. Sci. Health 1993, B28, 687-699.
Sovocool, G. W.; Lewis, R. G.; Harless, R. L.; Wilson, N. K.;
Zehr, R. Z. Analysis of technical chlordane by gas chroma-
tography/mass spectrometry. Anal. Chem. 1977, 49, 734-
740.
1
995, 103, 690-694.
Kr a¨ mer, P. Biosensors for measuring pesticide residues in the
environment: past, present, and future. J . AOAC Intl. 1996,
7
9, 1245-1254.
Langone, J . L.; Van Vunakis, H. Radioimmunoassay for
dieldrin and aldrin. Res. Commun. Chem. Pathol. Pharma-
col. 1975, 10, 163-171.
Ligler, F. S.; Golden, J . P.; Shriver-Lake, L. C.; Ogert, R. A.;
Wijesuria, D.; Anderson, G. P. Fiber-optic biosensors for the
detection of hazardous materials. Immunomethods 1993, 3,
Toppozada, A. R., Wright, J ., Eldefrawi, A. T; Eldefrawi, M.
E. Evaluation of a fiber optic immunosensor for quantitating
cocaine in coca leaf extracts. Biosens. Bioelectron. 1997, 12,
113-124.
1
22-127.
Lowry, O. H.; Rosebrough, N. J .; Farr, A. L.; Randall, R. J .
Protein measurement with the folin phenol reagent. J . Biol.
Chem. 1951, 193, 265-275.
Metcalf, R. L.; Kapoor, I. P.; Lu, P.; Schuth, C. K.; Sherman,
P. Model ecosystem studies of environmental fate of six
organochlorine pesticides. Environ. Health Perspect. 1973,
Tu, C. M.; Miles, J . R. W. Interactions between insecticides
and soil microbes. Residue Rev. 1976, 64, 17-65.
Ungnade, H. E.; McBee, E. T. The chemistry of perchlorocy-
clopentenes and cyclopentadienes. Chem. Rev. 1958, 58,
249-320.
Van Wijnen, J . H.; Stijkel, A. Health risk assessment of
residents living on harbor sludge. Int. Arch. Occup. Environ.
Health 1988, 61, 77-87.
Wigfield, Y. Y.; Grant, R. Evaluation of an immunoassay kit
for the detection of certain organochlorine (cyclodiene)
pesticide residues in apple, tomato, and lettuce. Bull.
Environ. Contam. Toxicol. 1992, 49, 342-347.
World Health Organization (WHO). In Environ. Health Cri-
teria 38 Hept; World Health Organization: Genova, 1984,
pp 9-11.
Wurster, C. F. Aldrin and dieldrin. Environment 1971, 13, 33-
45.
Zhao, C. Q.; Anis, N. A.; Rogers, K. R.; Kline, R. H.; Wright,
J .; Eldefrawi, A. T.; Eldefrawi, M. E. Fiber optic immuno-
sensor for polychlorinated biphenyls. J . Agric. Food Chem.
1995, 43, 2308-2315.
4
, 35-44.
Murphy, D. L. Bioaccumulation of chlordane by channel catfish
Ictalurus punctatus) and toxicity to a human liver epithelial
(
cell line. Ph.D. Dissertation, University of Maryland, 1995.
Pfaffenberger, B.; Hardt, I.; Huhnerfuss, H.; Konig, W. A.;
Rimkus, G.; Glaush, A.; Schurig, V.; Hahn, J . Enantiose-
lective degradation of alpha-hexachlorocyclohexane and
cyclodiene insecticides in roe-deer liver samples from dif-
ferent regions of Germany. Chemosphere 1994, 29, 1543-
1
554.
Pourfarzaneh, M.; White, G. W.; Landon, J .; Smith, D. S.
Cortisol directly determined in serum by fluoroimmunoas-
say with magnetizable solid phase. Clin. Chem. 1980, 26,
7
30-733.
Rimkus, G. and Wolf, M. Contamination of game in Schleswig-
Holstein. 2. Residues of dieldrin, heptachlor epoxide and
other cyclodiene insecticides in the liver fat of hares (Lepus
europaeus L.). [German] Z. Lebensm.-Unters.-Forsch. 1987,
Received for review February 10, 1997. Revised manuscript
X
received May 9, 1997. Accepted May 9, 1997. This study was
1
84, 308-312.
supported in part by U.S. Environmental Protection Agency
Rogers, K. R.; Valdes, J . J .; Eldefrawi, M. E. Acetylcholine
(EPA) Grant CR-820460 to M.E. and Fellowship U914418-01-1
receptor fiber-optic evanescent fluorosensor. Anal. Biochem.
awarded to K.E.B.
1
989, 182, 353-359.
J F9701190
Rogers, K. R.; Cao, C. J .; Valdes, J . J .; Eldefrawi, A. T.;
Eldefrawi, M. E. Acetylcholinesterase fiber optic biosensor
for detection of anticholinesterases. Fundam. Appl. Toxicol.
X
Abstract published in Advance ACS Abstracts, J uly,
1
991, 16, 810-820.
1, 1997.