Development of an Enzyme-Linked Immunosorbent Assay for Azadirachtins

Article abstract of DOI:10.1021/jf960688s

An enzyme-linked immunosorbent assay for azadirachtin A, a bioactive tetranor-triterpenoid from the neem tree (Azadirachta indica A. Juss), was developed. The conjugates were synthesized by employing the hemisuccinate method and an ester activation method, which were previously developed. Four rabbits were immunized with either 7-hemisuccinylazadirachtin-keyhole limpet hemocyanin (KLH) or 3-hemisuccinylazadirachtin-KLH. All sera were tested against the coating antigens 7-hemisuccinylazadirachtin-bovine serum albumin (BSA) and 3-hemisuccinylazadirachtin-BSA. From this screening procedure, an assay was developed using antiserum 7056 . The assay showed an IC₅₀ value of about 100 ppb, with a range of detection from 10 to 1000 ppb for azadirachtin A and related compounds. Thus, the method presented is a step forward in terms of inexpensive, fast, and sensitive (improvement by a factor of 100) analysis of azadirachtin and should provide a method that can easily be used more widely.

Full text of DOI:10.1021/jf960688s

J. Agric. Food Chem. 1997, 45, 2363
−2368
2363
Develop m en t of a n En zym e-Lin k ed Im m u n osor ben t Assa y for
Aza d ir a ch tin s†
Stefan Schu¨tz,‡ Ingrid Wengatz,§ Marvin H. Goodrow,§ Shirley J . Gee,§ Hans E. Hummel,‡ and
Bruce D. Hammock*,§
Institute of Phytopathology and Applied Zoology, Department of Biological and Biotechnical Plant
Protection, J ustus Liebig University, D-35390 Giessen, Germany, and Department of Entomology and
Environmental Toxicology, University of California, Davis, California 95616
An enzyme-linked immunosorbent assay for azadirachtin A, a bioactive tetranor-triterpenoid from
the neem tree (Azadirachta indica A. J uss), was developed. The conjugates were synthesized by
employing the hemisuccinate method and an ester activation method, which were previously
developed. Four rabbits were immunized with either 7-hemisuccinylazadirachtin-keyhole limpet
hemocyanin (KLH) or 3-hemisuccinylazadirachtin-KLH. All sera were tested against the coating
antigens 7-hemisuccinylazadirachtin-bovine serum albumin (BSA) and 3-hemisuccinylazadirachtin-
BSA. From this screening procedure, an assay was developed using antiserum 7056 . The assay
showed an IC50 value of about 100 ppb, with a range of detection from 10 to 1000 ppb for azadirachtin
A and related compounds. Thus, the method presented is a step forward in terms of inexpensive,
fast, and sensitive (improvement by a factor of 100) analysis of azadirachtin and should provide a
method that can easily be used more widely.
Keyw or d s: Azadirachtin; immunoassay; neem
INTRODUCTION
the design of antigens, the subsequent development of
antibodies, and the main characteristics of the immu-
noassay for azadirachtin A are presented in this paper.
Azadirachtin is a secondary plant product of the neem
tree (Azadirachta indica A. J uss). It is the main active
component of neem seeds and a prime example of
biorational insecticides that disturb an insect’s develop-
ment rather than the biochemical or metabolic activity
of enzymes that are found more ubiquitously in nature
(Schmutterer and Rembold, 1980). Because of its selec-
tive action against target pest insects (Schmutterer,
1988) and low mammalian toxicity (Schmutterer, 1990),
it is an insecticide of considerable environmental safety
(Schmutterer, 1995). Nevertheless, both ecotoxicological
studies and quality control of this natural product were
hindered by difficulties in setting up sufficiently sensi-
tive and inexpensive methods for analytical evaluation.
Current methods utilizing high-performance liquid chro-
matography-ultraviolet spectrophotometry (Schneider
and Ermel, 1987) are expensive and not acceptably
sensitive. Widely variable amounts of azadirachtin in
the biological raw material (neem kernels), depending
on origin and storage conditions, pose serious problems
regarding quality assurance of biorational pesticides
based on azadirachtin. However, quick monitoring of
large numbers of samples for product quality control and
environmental fate studies demands more cost-effective
screening methods with the least amount of sample
pretreatment. Immunochemical assays offer such a
desired screening method, because they are very sensi-
tive, often need little or no sample cleanup or prepara-
tion, and can be applied to numerous samples simulta-
neously (Karu et al., 1992). The synthetic strategy for
MATERIALS AND METHODS
Ch em ica ls. Azadirachtin A (compound 1; Figure 1) was
extracted from neem kernels (Schmutterer and Rembold, 1980)
and purified according to a procedure similar to a multilayer
countercurrent chromatography (MLCCC) method for mar-
rangin (Hein and Hummel, 1995) to a purity of approximately
90%. Immunochemicals were purchased from Sigma Chemical
Co. (St. Louis, MO). Other chemical reagents were obtained
from Aldrich Chemical Co. (Milwaukee, WI). Solvents for
synthesis were distilled and dried; solids were recrystallized
and dried prior to reaction. The borate buffer for protein
conjugation consisted of 1.9 g of Na2B4O7‚10H2O, 0.9 g of NaCl,
and 0.02 g of NaN3 in 100 mL of H2O, pH adjusted to 7.8 with
1 M HCl; 0.1 M PBS solution consisted of 8 g of NaCl, 1.15 g
of Na2HPO4, 0.2 g of KH2PO4, and 0.2 g of KCl in 100 mL of
H2O adjusted with HCl to pH 7.5. PBST consisted of 0.01 M
PBS with 0.05% Tween 20.
In str u m en ts. Normal phase thin-layer chromatograms
(TLC) were performed on 0.25 mm precoated silica gel F254
plastic sheets from Merck (Darmstadt, Germany) and evalu-
ated by iodine vapor staining. 1H nuclear magnetic resonance
(1H-NMR) spectra were obtained with
a Bruker QE-300
spectrometer (Billerica, MA) operating at 300.1 MHz. Chemi-
cal shifts (δ) are given relative to tetramethylsilane (TMS) as
an internal reference. Mass spectra were obtained with a
quadrapole mass spectrometer VG Trio-2 (VG Masslab, Al-
trincham, U.K.) by electron impact ionization (EI). Immu-
noassay absorbances were read with a Spectra Max 250
microplate reader (Molecular Devices, Menlo Park, CA) in
dual-wavelength mode (450-650 nm). The binding and
competition curves were analyzed with a commercial software
package (Softmax Pro 1.20, Molecular Devices) using a four-
parameter logistic equation. Unless otherwise mentioned, all
data presented from ELISA experiments correspond to the
average of at least three well replicates on one plate. Poly-
styrene microtiter plates were purchased from Nunc (Max-
isorb, Roskilde, Denmark, catalog no. 442404).
* Author to whom correspondence should be ad-
dressed [telephone (916) 752-7519; fax (916) 752-1537;
e-mail bdhammock@ucdavis.edu].
† Dedicated to Prof. H. Schmutterer on the occasion
of his 70th birthday on April 11, 1996.
‡ J ustus Liebig University.
P r ep a r a tion of Ha p ten s. Preparation of 7-Hemisuccinyl-
azadirachtin (2a ). Freshly distilled dry pyridine (8 mL) and
§ University of California.
S0021-8561(96)00688-7 CCC: $14.00
© 1997 American Chemical Society

2364 J. Agric. Food Chem., Vol. 45, No. 6, 1997
Schu¨tz et al.
1) of a white glassy product, which was characterized by NMR
and MS as 3-deacetylazadirachtin (2b). The bis(hemisucci-
nates) from succinylation of both the 3- and 7-hydroxy groups
had a much lower Rf on TLC. Thus, we were able to monitor
its appearance. Using the reaction conditions described, both
TLC and NMR indicate that none of the bis derivative was
formed. In solution, 2b was rather unstable and, therefore,
was used immediately in further synthesis (50 mg) and for
identification (13 mg). 1H NMR (CDCl3) δ 6.85 (dq, J d ) 1.0
Hz, J q ) 7.0 Hz, 1 H, 3′), 6.44 (d, J ) 2.9 Hz, 1 H, 23), 5.65 (s,
1 H, 21), 5.04 (d, J ) 2.9 Hz, 1 H, 22), 5.02 (s, 1 H, 11-OH),
4.74 (s, 1 H, 7), 4.65 (d, J ) 3.3 Hz, 1 H, 15), 4.59 (dd, J ) 2.6,
12.4 Hz, 1 H, 6), 4.21 (d, J ) 10.0 Hz, 1 H, 19a), 4.05 (d, J )
8.7 Hz, 1 H, 28b), 3.76 (s, 3 H, methyl ester), 3.69 (s, 3 H,
methyl ester), 3.60 (d, J ) 9.9 Hz, 1 H, 19b), 3.20 (d, J ) 12.4
Hz, 1 H, 5), 3.00 (br, 1 H, 7-OH), 2.75 (s, 1 H, 20-OH), 2.37 (d,
J ) 5.2 Hz, 1 H, 17), 2.1 (br, 1 H, 3-OH), 2.00 (s, 3 H, 18), 1.85
(s, 3 H, 5′), 1.78 (d, J ) 7.0 Hz, 3 H, 4′), 1.75 (s, 3 H, 30); EI-
MS (70 eV), m/e (rel intensity) 95 (100), 151 (54), 195 (4.5),
229 (4.5), 277 (3.3), 347 (3.7), 403 (2.1), 475 (1.7), 507 (1.3),
517 (11.7), 521 (1.3), 535 (0.7), 546 (1.9), 561 (1.5), 579 (1.7),
601 (0.6), 619 (0.2) (no parent ion at 677).
Preparation of 3-Hemisuccinylazadirachtin (3b). The prepa-
ration of this compound was done in analogy to the preparation
of compound 2a . Freshly distilled dry pyridine, 8 mL, and 40
mg of dry (5 h, 0.03 Torr) 2b were mixed under nitrogen
flushing. Dry (5 h, 0.03 Torr) succinic anhydride (100 mg),
followed by 50 µL of diisopropylcarbodiimide and a few crystals
of DMAP, was transferred into the nitrogen-flushed solution.
The mixture was stirred overnight at room temperature in the
dark, yielding 197 mg of a light yellow solid after separation
of the precipitate and evaporation of the solvent. Analytical
TLC showed only one additional spot at Rf ) 0.06 (solvent
system 1); preparative TLC yielded 31 mg (68% with respect
to 2b) of a white glassy product, which was identified by NMR
as 3-hemisuccinylazadirachtin 3b. The bis(hemisuccinate)
from succinylation of both the 3- and 7-hydroxy groups had a
much lower Rf on TLC. Thus, we were able to monitor its
appearance. Using the reaction conditions described, both
TLC and NMR indicate that none of the bis derivative was
formed. The formation of detectable amounts of 2a was
avoided by optimization of reaction time. In TLC solvent and
NMR solvent systems, 3b was rather unstable and, therefore,
was used immediately for further synthesis (16 mg) and
characterization (15 mg). The degradation half-life time in
DMSO-d6 solvent at 20 °C was about 20 min and at -20 °C
F igu r e 1. Synthesis of haptens 2a and 3b used for immuno-
gens and for coating antigens.
60 mg of dry 1 (5 h, 0.03 Torr) were mixed under nitrogen
flushing. Dry (5 h, 0.03 Torr) succinic anhydride (100 mg),
followed by 50 µL of diisopropylcarbodiimide (99%) and a few
crystals of 4-(dimethylamino)pyridine (DMAP), was trans-
ferred with the aid of a nitrogen hood into the nitrogen-flushed
solution. The mixture was stirred overnight at room temper-
ature in the dark. The pyridine solvent was separated from
the precipitated diisopropylurea and was evaporated (4 h, 0.03
Torr) to yield 173 mg of a yellow oil. Analytical TLC (solvent
system 1, toluene/ethyl acetate/methanol, 5:1:2) showed only
one additional spot at Rf ) 0.08; preparative TLC yielded 30
mg (44% with respect to 1) of a white glassy product, which
was identified by NMR as 3-hemisuccinylazadirachtin (2a ).
This compound was rather unstable and, therefore, was used
immediately in further synthesis (15 mg) and for identification
(15 mg). The degradation half-life time in solvents such as
CDCl3 and DMSO-d6 at 20 °C was about 30 min and at -20
°C about 2 h (kinetics determined by TLC). Therefore, only
1H-NMR (DMSO-d6) spectra of mixtures of the parent com-
pound and the degradation products could be obtained. The
esterification of the free 7-OH moiety was concluded from the
1
about 2 h (kinetics determined by TLC). Therefore, only H-
NMR (DMSO-d6) spectra of mixtures of the parent compound
and the degradation products could be obtained. The esteri-
fication of the free 3-OH moiety was concluded from the
1
disappearance of the broad peak in H-NMR (DMSO-d6) at 2.07
ppm (br s, 1 H, 3-OH) and the observation of a new peak at
2.65 ppm (br s, 4 H, (CH2)2) and of a peak shift from 4.84 (s,
1 H, 7) to 5.36 ppm (s, 1 H, 7).
P r ep a r a tion of Ha p ten -P r otein Con ju ga tes. Haptens
2a and 3b were conjugated to bovine serum albumin (BSA),
keyhole limpet hemocyanin (KLH), and tyramine. Tyramine
conjugate formation was monitored by TLC to optimize the
conjugation conditions. Spectrophotometric examination of the
conjugation products was not possible because of the weak
absorbance of the haptens in the UV-vis range.
Preparation of the Activated 7-Hemisuccinylazadirachtin
(3a ). A three-neck, 25-mL round-bottom flask with a magnetic
stirring bar, two rubber septa, and one glass stopper was
flushed with nitrogen. Anhydrous tetrahydrofuran (THF, 5
mL) and 15 mg of dry (5 h, 0.03 Torr) 2a were mixed under
nitrogen flushing. Dry (5 h, 0.03 Torr) N-hydroxysuccinimide
(10 mg) was transferred with the aid of a nitrogen hood into
the nitrogen-flushed solution. The mixture was stirred over-
night on ice in the dark. The solvent was evaporated in a
stream of nitrogen to yield 24 mg of a pale yellow solid that
showed one additional spot in TLC with Rf ) 0.67 (solvent
system 1). Preparative TLC yielded 10 mg (59% with respect
2a ) of 3a , as a white solid. The unstable and highly moisture
sensitive compound was used immediately for further synthe-
sis.
1
disappearance of the broad peak in H-NMR (DMSO-d6) at 3.0
ppm (br s, 1 H, 7-OH) and the observation of a new peak at
2.65 ppm [br s, 4 H, (CH2)2] and of a peak shift from 4.75 ppm
(s, 1 H, 7) to 5.26 ppm (s, 1 H, 7).
Preparation of 3-Deacetylazadirachtin (2b). A three-neck,
25-mL round-bottom flask with a magnetic stirring bar, two
rubber septa, and one glass stopper was flushed with nitrogen.
Anhydrous methanol (10 mL) and 200 mg of dry (5 h, 0.03
Torr) azadirachtin (1) were mixed under a stream of nitrogen.
Sodium hydride (120 mg; 60% in mineral oil) was transferred
with the aid of a nitrogen hood into the nitrogen-flushed
azadirachtin solution. The mixture was allowed to stir for 30
min in the dark. Thereafter, excess of NH4Cl was added and
the solvent was evaporated in a stream of nitrogen. The
residue was resuspended in dry CH2Cl2 and filtered. Evapora-
tion of the filtrate yielded 220 mg of a yellow-orange oil
showing four spots on analytical TLC. Preparative TLC (Rf
) 0.22, solvent system 1) yielded 63 mg (32% with respect to

ELISA for Azadirachtins
J. Agric. Food Chem., Vol. 45, No. 6, 1997 2365
Preparation of the Activated 3-Hemisuccinylazadirachtin
(4b). Dry (5 h, 0.03 Torr) 3b (16 mg) was treated analogously
to the preparation procedure of 3a . The solvent was evapo-
rated in a stream of nitrogen to yield 27 mg of a pale yellow
solid that showed one additional spot in the analytical TLC
with Rf ) 0.72 (solvent system 1). Preparative TLC yielded
10 mg (56% with respect to 3b) of 4b, a white solid. The
unstable and highly moisture sensitive compound was used
immediately for further synthesis.
repeated every 4 weeks. Animals were bled (2-5 mL) 10 days
after each boost from the central ear artery under light
anesthesia. All procedures, except for the preparation of the
immunogens, were carried out by personnel of the Laboratory
Animal Health Clinic at the University of California at Davis.
The blood was kept overnight at 4 °C to allow clotting. After
centrifugation (500g, 10 min), removal of the clot, and a second
centrifugation (500g, 10 min), the clear serum was carefully
removed with
a Pasteur pipet, aliquoted, and stored in
cryovials at -20 °C.
Preparation of the BSA Conjugates. These reactions were
conducted in 5-mL glass flasks containing magnetic stirring
bars. BSA (fraction V, 96-99%) was dissolved (10 mg) in 2
mL of 0.2 M borate buffer, and 0.2 mL of dimethylformamide
(DMF) was added very slowly (ca. 15 min) with rigorous
stirring at room temperature, causing a slight precipitation.
A solution of 5 mg of active ester (3a or 4b) in 0.2 mL of
anhydrous DMF was added slowly (ca. 20 min). A small
amount of precipitate developed during this addition. The
mixture was stirred gently overnight on ice and then dialyzed
versus 0.2 M PBS (five changes of 2 L each, 5 days, at room
temperature) in a SpectraPor membrane (6000-8000 MW
cutoff) to desorb and separate hapten molecules that were
bound to the protein by physisorption only. The dialyzed
solution was transferred into polypropylene vials and stored
at -30 °C.
Scr een in g of An tiser a w ith Tw o-Dim en sion a l Titr a -
tion . For the determination of initial titers of each serum, a
microtiter plate (Nunc) was coated with 100 µL/well of a
dilution series of coating antigen (0.016-2 µg/mL) and an
additional row with an indifferent fenvalerate-conjugated BSA
antigen (2 µg/mL) in 0.5 M carbonate buffer (pH 8.5). Plates
were sealed with adhesive plate sealers and incubated over-
night at 4 °C. The following day, the coated plates were
washed three times with PBST. Diluted serum (100 µL/well)
was added to each well and incubated for 2 h at room
temperature. The sera were checked in a two-dimensional
titration dilution series (1:500 to 1:256000 and two controls
containing a fenvalerate antiserum 1:20000 in 0.2 M PBS, pH
7.5) against the dilution series of coating antigens. After
another washing step, 100 µL/well of goat anti-rabbit IgG
conjugated to horseradish peroxidase (HRP; GAR/HRP), di-
luted 1:8000 in 0.2 M PBS, pH 7.5, was added and incubated
for 2 h at room temperature. The plates were washed again,
and 100 µL/well of substrate solution [800 µL of a 1.2%
3,3′,5,5′-tetramethylbenzidine (TMB) in DMSO and 13.2 µL
of 30% H2O2 in 100 mL of a 0.1 M acetate buffer, pH 5.5] was
added. After 10-20 min, the enzyme reaction was stopped
with 50 µL/well of 4 M H2SO4. The absorbance was read at
450-650 nm.
Preparation of the KLH Conjugates. The reactions were
conducted analogous to the preparation of BSA conjugates.
Precipitate developed during the addition of the active ester
(3a or 4b). The dialyzed solution was transferred into
polypropylene vials and stored at -30 °C.
Preparation of Tyramine Derivatives. The reaction was
conducted in a 5-mL glass flask containing a magnetic Teflon
stirring bar. Tyramine hydrochloride (99%), 10 mg, was
dissolved in 2 mL of 0.2 M borate buffer. DMF (0.2 mL) was
added very slowly (ca. 15 min) during rigorous stirring at room
temperature, causing a slight precipitation. The active ester
(3a or 4b) was dissolved (5 mg) in 0.2 mL of anhydrous DMF
and added very slowly (ca. 20 min). A slight precipitate
developed during this addition. The solution was stirred
gently overnight on ice, and the solvent was subsequently
evaporated (24 h, 0.03 Torr). The residue was redissolved in
2-propanol and subjected to preparative TLC with Rf ) 0.22
(solvent system 2, toluene/1-propanol, 5:1), yielding 1.4 mg
(25% with respect to the active ester) of the 7-hemisuccinyl-
azadirachtin-tyramine derivative. The yield of the 3-hemisuc-
cinylazadirachtin-tyramine derivative was only 0.8 mg (14%
with respect to the active ester). Therefore, the similar Rf
value on TLC examination with Rf ) 0.23 (solvent system 2)
was considered as tentative identification; 7-hemisuccinylaza-
dirachtin-tyramine was characterized by NMR and MS: 1H
NMR (CDCl3) δ 7.05 (m, 4H, aromatic), 6.90 (dq, J d ) 1.0 Hz,
J q ) 7.0 Hz, 1H, 3′), 6.82 (dd, J ) 2.1, 9 Hz, 4H, 7′), 6.60 (br,
1H, NH), 6.46 (d, J ) 2.9 Hz, 1H, 23), 5.65 (s, 1H, 21), 5.05 (d,
J ) 2.9 Hz, 1H, 22), 5.02 (s, 1H, 11-OH), 4.76 (s, 1H, 7), 4.65
(d, J ) 3.3 Hz, 1H, 15), 4.61 (dd, J ) 2.4, 12.6 Hz, 1H, 6), 4.17
(d, J ) 9.5 Hz, 1H, 19a), 4.05 (d, J ) 8.9 Hz, 1H, 28b), 3.76 (s,
3H, methyl ester), 3.68 (s, 3H, methyl ester), 3.64 (d, J ) 9.8
Hz, 1H, 19b), 3.30 (d, J ) 12.4 Hz, 2H, 5, 9), 3.2 (br, 1H,
tyramine-OH) 2.9 (br, 1H, 20-OH), 2.37 (d, J ) 5.2 Hz, 1H,
17), 2.03 (s, 3H, 18), 1.85 (s, 3H, 5′), 1.78 (d, J ) 7.0 Hz, 3H,
4′), 1.75 (s, 3H, 30); EI-MS (70 eV), m/e (rel intensity) 95 (90),
105 (100), 120 (65), 143 (23), 151 (54), 195 (4.5), 229 (4.0), 277
(3.1), 347 (3.4), 403 (2.1), 475 (1.8), 507 (1.3), 521 (1.2), 535
(1.4), 545 (1.3), 559 (3.7), 595 (9.5), 588 (1.1), 620 (0.8), 661
(0.3), 704 (0.4), 719 (0.1) (no parent ion at 955).
ELISA Coa tin g An tigen F or m a t. Com p etition . To
determine the analytical characteristics of the ELISA coating
antigen format for each serum, one microtiter plate was
divided into two equal (12 columns, 3 rows) parts and one
unequal (12 columns, 2 rows) part. The equal parts were
coated with 100 µL/well of the two coating antigens 2a -BSA
and 3b-BSA, and the unequal part was coated with
a
fenvalerate-BSA conjugate as a protein control, with the
optimized concentrations (0.03 µg/mL 2a -BSA, 1 µg/mL 3b-
BSA, 1 µg/mL fenvalerate-BSA) overnight at 4 °C. The next
day, the coated plates were washed three times with PBST
and the analyte (0.0013-100 µg/mL in 0.2 M PBS, pH 7.5, 50
µL/well) and 50 µL/well of the optimized dilution of the
antiserum (no. 7053 or 7056, 1:20000; no. 7052 or 7054,
1:100000 in 0.2 M PBS, pH 7.5) were added and incubated for
2 h at room temperature. After another washing step, 100
µL/well of GAR/HRP, diluted 1:8000 in 0.2 M PBS, pH 7.5,
was added and incubated for 1 h at room temperature. The
plates were washed again, and 100 µL/well of substrate
solution was added; after 10-20 min, the enzyme reaction was
stopped with 50 µL/well 4 M H2SO4. The absorbance was read
at 450-650 nm.
RESULTS AND DISCUSSION
Ha p ten Syn th esis. Since antibodies recognize best
the part of the hapten that is most distant from the
conjugate linkage, a carboxylic group opposite from the
most characteristic groups of the molecule was desir-
able. However, the complex structure of azadirachtin
with a densely packed array of oxygen functionalities
of many different types made it difficult to determine
which portions of the molecule are most unique. In the
course of studies on azadirachtin, it was determined
(Simmonds et al., 1995a) that there is not much change
in the biological activity of azadirachtin when the acyl
groups attached to the 1- and 3-hydroxys are modified
or removed. It is known from extensive synthetic work
(Grossmann and Ley, 1994) on azadirachtin that the
3-hydroxy group of azadirachtin is readily saponified
Im m u n iza tion of Ra bbits. Rabbits (female New Zealand,
white, 3-5 kg) 7052 and 7054 were immunized with 2a -KLH,
and rabbits 7053 and 7056 were immunized with 3b-KLH.
Rabbits were injected intradermally (5-20 sites per animal)
with 100 µg of the corresponding hapten-protein conjugate.
For injection, the immunogens were dissolved in 0.9% sterile
NaCl solution and mixed 1:1 with Freund’s complete adjuvant
to a final volume of 500 µL/injection. For the boost immuniza-
tion, the immunogens were prepared in the same way except
Freund’s incomplete adjuvant was used. The boosting was
started 1 month after the initial immunization and was

2366 J. Agric. Food Chem., Vol. 45, No. 6, 1997
Schu¨tz et al.
and re-esterified in good yield without concomitant
rearrangement of the azadirachtin skeleton. Therefore,
the 3-position was chosen for attachment of the handle
to provide optimization of recognition and the 7-position
for reason of simplicity because no saponification prior
to derivatization was necessary at the free 7-hydroxy
group.
For reasons of literature precedence (Wing et al.,
1978) and commercial availability, succinic acid was
chosen as a linker to join azadirachtin to the carrier
protein. Since azadirachtin is a comparably large
molecule, the succinate derivatization should not present
a serious handle recognition problem. Ester formation
was used as a means of linking because it is a mimic of
the natural derivatives, especially the 3′-acyl compound.
Because of the short linker group, the attaching ester
bond should be shielded by the large azadirachtin
molecule from ester hydrolase-catalyzed hydrolysis.
Purification problems due to insufficient separation
of azadirachtin-hemisuccinate from succinic acid by
TLC could be solved by switching the solvent from THF
(high solubility of succinic acid) to pyridine (low solubil-
ity for succinic acid). When the deacetylated 2b was
reacted with hemisuccinate, we initially expected a
mixture of succinyl groups at both the 3- and 7-positions.
However, these monosuccinylated derivatives were eas-
ily distinguished by their Rf on TLC. TLC thus indi-
cated that under the reaction conditions used, no
disuccinate was formed and no detectable 7-hemisucci-
nylazadirachtin was present in the 3-hemisuccinylaza-
dirachtin preparation. This chromatographic evidence
was supported by NMR data. Although this observation
was initially surprising, the ease with which the 3-posi-
tion of azadirachtin is deacylated and selectively reacy-
lated (Grossmann and Ley, 1994) indicates that this
position is sufficiently reactive to allow a selective
reaction to occur without acetylating the other second-
ary hydroxy group. The stability of the hapten molecule
is low, because of the destabilizing effect of the free
carboxylic group. The intrinsic acidity of the compound
may be causing rapid degradation of the hapten mol-
ecule. The short chain length of the spacer was chosen
to reduce intramolecular reactions. An approach using
an ω-amino acid as linker (Grossmann and Ley, 1994)
might yield more stable haptens.
Con ju ga tion of Ha p ten s to P r otein s a n d Con fir -
m a tion of Con ju ga tion . Activation of the carboxylic
acid by N-hydroxysuccinimide was utilized to suppress
de- and re-esterifications within the azadirachtin mol-
ecule. Using a low reaction temperature significantly
decreased the amount of byproducts. Solubility prob-
lems of either azadirachtin or protein were solved by
employing a DMF-water mixture as a solvent for
performing hapten-protein conjugation (Karu et al.,
1992). An estimate of hapten density could not be
achieved by spectrophotometric methods, since the
absorption of the azadirachtin molecule and the handle
is weak within the UV-vis range. Therefore, we used
tyramine as a mimic of protein NH2 moieties. This led
to an estimate of 0.6 mg of hapten (20% yield)/10 mg of
protein. Using a molecular mass of 68 000 amu for BSA,
a hapten density of 4 per BSA-molecule was calculated.
The same quantity of tyramine hydrochloride (10 mg)
was used. Because of the lower molecular mass of
tyramine-HCl, however, tyramine was in a molar
excess compared to BSA. Thus, tyramine conjugate
yields may be elevated. Moreover, not every amino acid
of BSA or KLH is capable of binding haptens. There-
F igu r e 2. Antibody binding of conjugated azadirachtin.
ELISA was used to evaluate the binding of antibodies raised
against 2a -KLH to 96-well microplates coated with fenval-
erate-conjugated BSA or the conjugates 2a -BSA or 3b-BSA.
Microplates coated with the conjugates were exposed to various
dilutions of antiserum (concentration × 1000). Similar experi-
ments were repeated for other antigen-antibody combinations
to determine effective antibody titers.
fore, the estimate of four hapten molecules attached to
a protein molecule might be too high. An approach
using a handle with dye properties (Sashidar et al.,
1994) would facilitate an estimate of the success of
conjugation.
ELISA Resu lts: Recogn ition of Con ju ga ted Aza -
d ir a ch tin . Antibodies raised against KLH conjugates
of haptens 2a and 3b were tested by ELISA for binding
to conjugates of the same haptens to BSA. Figure 2
shows the binding of one selected antiserum (anti-2a -
KLH) to 2a -BSA, 3b-BSA and a fenvalerate-conju-
gated BSA as a negative control, confirming the covalent
attachment of 2a and 3b to BSA and demonstrating the
ability of this antiserum to recognize the azadirachtin
structure. The superior binding of 2a -BSA (in prefer-
ence to 3b-BSA) shown in Figure 2 was typical of the
antisera from rabbits immunized with conjugates of
hapten 2a , suggesting significantly better binding to
homologous conjugates (spacer attached at the same
point on the azadirachtin molecule as the immunizing
hapten) than to heterologous conjugates (spacer at-
tached at a different site). The above data suggest that
these antisera can discriminate between positions of
conjugation to the azadirachtin skeleton and that
homology of spacer attachment may be an important
determinant of binding strength. However, because of
the uncertainty of the exact nature of hapten-protein
conjugates, specificity studies are most conclusive when
conducted with free haptens in a competitive immu-
noassay, as described in the next section.
In d ir ect Com p etitive ELISA Resu lts: Recogn i-
tion of F r ee Aza d ir a ch tin . Antibodies were tested
for binding of several azadirachtin-related compounds
(Figure 3), including azadirachtin A, deacetylazadirach-
tin, and several methanol hydrolysates of azadirachtin
by using indirect competitive ELISA. However, only the
combination of 2a -BSA as coating antigen and antibody
7056 yielded good competition curves with azadirachtin
A. In Table 1, azadirachtin A and methanol-hydrolyzed
azadirachtin are shown to compete for the binding to
the antibody with conjugated 7-hemisuccinylazadirach-
tin. Table 1 data indicate that this antibody recognizes
the target compound azadirachtin and related com-
pounds (3-deacetylazadirachtin, 1-detigloylazadirachtin,
1-detigloyl-3-deacetylazadirachtin, azadirachtinin), not
just their hemisuccinyl derivatives. This illustrates the

ELISA for Azadirachtins
J. Agric. Food Chem., Vol. 45, No. 6, 1997 2367
they fail in detecting bioactive degradation products of
azadirachtin A because of missing UV-absorbing moi-
eties such as the tigloyl or the vinyl ether group. The
ELISA for azadirachtin A and related compounds has
detection limits of about 10 ng/g and is able to detect
these early products of azadirachtin A hydrolysis. The
bioactivity of the identified hydrolysis products of aza-
dirachtin was examined recently (Simmonds et al.,
1995b).
The usefulness of this immunoassay depends on its
ability to solve two different problems. One problem is
to discriminate compounds of different biological activity
to facilitate quality assurance of biorational insecticides
based on azadirachtins. Evidence is presented in this
study that the ELISA can discriminate among the tested
compounds. Extended screening of bioactive limonoids
will be a subject of following papers using both chro-
matographic separation (Schneider and Ermel, 1987;
Hein and Hummel, 1995; Kra¨mer et al., 1994) and
multianalyte approaches (Wortberg et al., 1996). The
relative importance of structural changes in recognition
of azadirachtin-related compounds, and practical utili-
zation of the developed assays will be investigated
further.
The second problem is the need for quick monitoring
of large numbers of samples for environmental fate
studies of azadirachtin that demand cost-effective screen-
ing methods with the least amount of sample pretreat-
ment. This ELISA based on antibodies against aza-
dirachtin-related compounds will be a useful tool for
screening requirements, especially because of its unri-
valed sensitivity. Levels of azadirachtin detection in
environmental samples such as soil (Stark and Walter,
1995) or tree foliage (Sundaram and Curry, 1994) have
been in the parts per million (µg/g) range, whereas the
presented ELISA method allows detection in the parts
per billion (ng/g) range. Thus, the method presented is
a step forward in terms of inexpensive, fast, and
sensitive analysis of azadirachtins and could therefore
have wide application.
F igu r e 3. Structures of azadirachtin derivatives tested for
cross-reactivity.
Ta ble 1. Sta n d a r d Cu r ve for F r ee Aza d ir a ch tin A a n d
Hyd r olyzed Aza d ir a ch tin Usin g 7056 (1:40000) An tiser u m
a n d 2a -BSA Con ju ga te (0.2 m g/L) a s Coa tin g An tigen
IC50
compound
azadirachtin A
3-deacetylazadirachtin
1-detigloylazadirachtinb
1-detigloyl-3-deacetylazadirachtinb
azadirachtinb
no. (µg/L) SD CRa (%)
1
100
60
41
35
50
45
42
100
63
2b
200
220
250
50
45
40
a
CR ) (IC50 for azadirachtin A/IC50 for compound) × 100.
b
Shown in Figure 3.
potential usefulness of these antibodies for detection and
measurement of azadirachtin-related compounds at
levels commonly found in environmental samples. The
difference in IC50 confirms the ability of this serum to
distinguish among closely related structures on the
basis of minor changes (i.e. in an acyl group).
ABBREVIATIONS USED
Assa y Ch a r a cter istics. Under optimized conditions
(3b-KLH antiserum 7056 dilution 1:40000 final dilu-
tion in the well, 2a -BSA 0.2 µg/mL) the assay for
azadirachtin A showed a standard curve that allowed
the measurement of the analyte from 10 to 1000 µg/L
with an average IC50 value of 100 ( 41 µg/L (Table 1).
The assay showed comparable sensitivity toward aza-
dirachtin-related compounds such as deacetylazadirach-
tin and other early products of azadirachtin hydrolysis.
The analyte concentration in this study is the original
concentration pipetted into the wells of the microtiter
plates. This was preferred because of better comparison
with conventional methods which also refer to the
original concentration in the sample.
BSA, bovine serum albumin; CR, cross-reactivity )
(IC50 for azadirachtin A/IC50 for compound) × 100;
DMAP, 4-(dimethylamino)pyridine; DMF, dimethylfor-
mamide; DMSO, dimethyl sulfoxide; ELISA, enzyme-
linked immunosorbent assay; GC, gas chromatography;
HPLC, high-performance liquid chromatography; HRP,
horseradish peroxidase; IC50, concentration of the ana-
lyte, where 50% of the antibody is inhibited by the
analyte, midpoint of the assay; KLH, keyhole limpet
hemocyanin; OD, optical density; OVA, ovalbumin; PBS,
phosphate-buffered saline; PBST, phosphate-buffered
saline with 0.5% Tween 20 (polyoxyethylene sorbitan
monolaurate); TLC, thin-layer chromatography; TMB,
3,3′,5,5′-tetramethylbenzidine.
CONCLUSION
ACKNOWLEDGMENT
Ecotoxicological studies and quality control of aza-
dirachtin have been hindered by difficulties in setting
up sufficiently sensitive and inexpensive analytical
evaluation methods. Widely variable amounts of aza-
dirachtin in the biological raw material (neem kernels),
depending on origin and storage conditions, pose serious
problems regarding quality assurance of biorational
insecticides based on azadirachtin. Conventional HPLC
methods using UV detection methods are able to detect
azadirachtin A down to 0.2 µg/g (Hull et al., 1993), but
S.S. was partially supported by a scholarship from
DFG (Deutsche Forschungsgemeinschaft, Bonn, Ger-
many) Grant SCHU 1135/1-1. The azadirachtin A was
purified in the laboratory of Dr. Y. Ito, National
Institutes of Health, Bethesda, MD. Further purifica-
tion work was performed by D. Hein, H. Vedder, and
Dr. Y. Ma. Support of the EU commission to D. Hein
is gratefully acknowledged (Grant AIR 2-CT 94-1343).
We thank Dr. A. Daniel J ones of the Facility for

2368 J. Agric. Food Chem., Vol. 45, No. 6, 1997
Schu¨tz et al.
3rd International Neem Conference (Nairobi, 1986); GTZ-
Verlag: Eschborn, 1987; pp 129-139.
Simmonds, M. S. J .; Blaney, W. M.; Grossmann, R. B.; Ley, S.
V. Behavioural responses of locusts and Spodoptora littoralis
to azadirachtin analogues containing fluorescent and im-
munogenic reporter groups. J . Insect Physiol. 1995a , 41,
555-564.
Simmonds, M. S. J .; Blaney, W. M.; Grossmann, R. B.; Ley, S.
V.; Anderson, J . C.; Banteli, R.; Denholm, A. A.; Green, P.
C. W.; Grossman, R. B.; Gutteridge, C.; J ennens, L.; Smith,
S. C.; Toogood, P. L.; Wood, A. behavioural and neurophysi-
ological responses of Spodoptora littoralis to azadirachtin
and a range of synthetic analogues. Entomol. Exp. Appl.
1995b, 77, 69-80.
Stark, J . D.; Walter, J . F. Persistence of azadirachtin A and B
in soil: effects of temperature and microbial activity. J .
Environ. Sci. Health, Part B 1995, 30, 685-698.
Sundaram, K. M. S.; Curry, J . Initial deposits and persistence
of azadirachtin in fir and oak foliage after spray application
of Margosan-O formulation. Pestic. Sci. 1994, 41, 129-138.
Wing, K. W.; Hammock, B. D.; Wustner, D. A. Development
of an S-bioallethrin specific antibody. J . Agric. Food Chem.
1978, 26, 1328-1333.
Wortberg, M.; J ones, G.; Kreissig, S.; 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, 291-303.
Advanced Instrumentation (U.C. Davis) for the mass
spectra and the Department of Chemistry (U.C. Davis)
for the use of their NMR instrument.
LITERATURE CITED
Grossmann, R. B.; Ley, S. V. Chemistry of insect antifeedants
from Azadirachta indica: synthesis of several derivatives
of azadirachtin containing fluorescent or immunogenic
reporter groups. Tetrahedron 1994, 50, 8871-8884.
Hein, D.; Hummel, H. E. Biologische Aktivita¨t gegenstrom-
chromatographisch gereinigter Extrakte aus Azadirachta
exelsa im Biotest mit Epilachna varivestis Muls. Mitt. Dtsch.
Ges. Allg. Angew. Entomol. 1995, 10, 247-250.
Hull, C. J ., J r.; Dutton, W. R.; Switzer, B. S. Quantification of
azadirachtins in insecticidal formulations by high-perfor-
mance liquid chromatography. J . Chromatogr. 1993, 633,
300-304.
Karu, A. E.; Goodrow, M. H.; Schmidt, D. J .; Hammock, B.
D.; Bigelow, M. W. Synthesis of haptens and derivation of
monoclonal antibodies for immunoassay of the phenylurea
herbicide diuron. J . Agric. Food Chem. 1992, 40, 301-309.
Kra¨mer, P. M.; Li, Q. X.; Hammock, B. D. Integration of liquid
chromatography with immunoassay: an approach combin-
ing the strengths of both methods. J . AOAC Int. 1994, 77,
1275-1287.
Sashidar, R. B.; Capoor, A. K.; Ramana, D. J . An improved
colorimetric method for quantification of protein conjugates.
J . Immunol. Methods 1994, 167, 121-127.
Schmutterer, H. Potential of azadirachtin containing pesticides
for integrated pest control in developing and industrialized
countries. J . Insect Physiol. 1988, 34, 713-719.
Schmutterer, H. Properties and potential of natural pesticides
from the neem tree, Azadirachta indica. Annu. Rev. Ento-
mol. 1990, 35, 271-297.
Schmutterer, H., Ed. The Neem Tree Azadirachta indica (A.
J uss) and Other Meliaceous Plants; Verlag Chemie: Wein-
heim, 1995.
Schmutterer, H.; Rembold, H. Zur Wirkung einiger Reinfrak-
tionen aus Samen von Azadirachta indica auf die Frassak-
tivita¨t und Metamorphose von Epilachna varivestis (Col.
Coccinellidae). Z. Angew. Entomol. 1980, 89, 179-188.
Schneider, B.; Ermel, K. Quantitative determination of aza-
dirachtin in crude and enriched extracts. In Proceedings,
Received September 9, 1996. Revised February 18, 1997.
Accepted March 5, 1997.X Partial support came from the
National Institutes Environmental Health Sciences Superfund
Basic Research Program 5 P42 PHS ES04699. U.C. Davis is
an NIEHS Health Sciences Center (ES05707) and an Envi-
ronmental Protection Agency Ecological Health Center (CR
819658). Although portions of this work were supported by
the U.S. EPA, no endorsement of the data is implied.
J F960688S
X Abstract published in Advance ACS Abstracts, May
1, 1997.