Evaluation of the carbonyl metallo immunoassay (CMIA) for the determination of traces of the herbicide atrazine

Article abstract of DOI:10.1016/S0022-328X(02)02098-3

The non-isotopic immunoassay (CMIA) was investigated for the quantification of the herbicide atrazine. This assay combines transition metal carbonyl complex and Fourier transform infrared spectroscopy. We describe the synthesis of three dicobalt hexacarbonyl tracers, derivatives of atrazine. Their relative binding affinity towards two purified IgG fractions of polyclonal rabbit anti-sera (anti-atrazine and -simazine) was evaluated by ELISA and CMIA. The best tracer 9 (bearing the longest alkyl chain between atrazine aromatic ring and metal carbonyl complex) was used for further study by CMIA. Reproducible competitive standard curves were obtained by using anti-simazine antibodies in PBS buffer pH 7.4 with pre-incubation of various amounts of atrazine and antibodies for 1 h at 4 °C before addition of the tracer. This study demonstrates the feasability of using CMIA for pesticide determination, although the sensitivity of current CMIA format does not reach the 0.1 μg 1^-1 level required by E.U.

Full text of DOI:10.1016/S0022-328X(02)02098-3

Journal of Organometallic Chemistry 668 (2003) 59Á66
/
www.elsevier.com/locate/jorganchem
Evaluation of the carbonyl metallo immunoassay (CMIA) for the
determination of traces of the herbicide atrazine
Nathalie Fischer-Durand, Anne Vessie`res *, Jan-Martin Heldt, Franck le Bideau,
Ge´rard Jaouen
Ecole Nationale Supe´rieure de Chimie de Paris, Laboratoire de Chimie organome´tallique, UMR CNRS 7576, 11 rue P. et M. Curie, 75231 Paris Cedex
05, France
Received 15 August 2002
Abstract
The non-isotopic immunoassay (CMIA) was investigated for the quantification of the herbicide atrazine. This assay combines
transition metal carbonyl complex and Fourier transform infrared spectroscopy. We describe the synthesis of three dicobalt
hexacarbonyl tracers, derivatives of atrazine. Their relative binding affinity towards two purified IgG fractions of polyclonal rabbit
anti-sera (anti-atrazine and -simazine) was evaluated by ELISA and CMIA. The best tracer 9 (bearing the longest alkyl chain
between atrazine aromatic ring and metal carbonyl complex) was used for further study by CMIA. Reproducible competitive
standard curves were obtained by using anti-simazine antibodies in PBS buffer pH 7.4 with pre-incubation of various amounts of
atrazine and antibodies for 1 h at 4 8C before addition of the tracer. This study demonstrates the feasability of using CMIA for
pesticide determination, although the sensitivity of current CMIA format does not reach the 0.1 mg lꢀ1 level required by E.U.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Atrazine; Metal carbonyl complex; Immunoassay; Fourier-transform infrared spectroscopy (FT-IR)
1. Introduction
to the complexity of water matrices. With the pioneer
development of antibodies production against small
molecules in the medical diagnostic domain, immuno-
chemical approach is now widely used in pesticide
monitoring: antibodies are designed to selectively trap
a given analyte or a group of structurally related
compounds in complex matrices. These immunochem-
ical methods are mostly based on immunoaffinity
chromatography (IAC) followed by on-line or off-line
Environmental protection is a growing concern today,
and government policies under public and health
organization pressures, are harsher towards residues
rates released in water and soil. In Europe, the council
directive 80/778/EEC relating to the quality of water
intended for human consumption fixed maximum allow-
able levels of pesticides in drinking water at 0.1 mg lꢀ1
for an individual pesticide and no more than 0.5 mg lꢀ1
for total pesticides in a given sample.
LC or LCÁ
/
MS quantification [3Á/6], and immunoassays
on microtitre plates [6Á
/
8] with detection by UVÁ
/vis
(ELISA technique: enzyme linked immunosorbent as-
say) or fluorescence (FIA: fluorescence immunoassay).
Our previous work in both the clinical diagnostic
Some sophisticated analytical methods are available
for the determination of residues in drinking or surface
waters, including gas chromatography coupled with
(carbonyl metallo immunoassay or CMIA) [9Á
/12] and
mass spectrometry (GCÁ
liquid chromatography coupled with MS (LCÁ
but they require cumbersome sample preparation, owing
/
MS) [1] and high performance
the environmental domain (IAC coupled with LC) [13Á
/
/MS) [2],
16], prompted us to evaluate the determination of
pesticides such as chlortoluron [17] and atrazine using
CMIA which uses metal-carbonyl complexes as tracers
and Fourier-transform infrared spectroscopy (FT-IR) as
the detection method. Atrazine, a member of the chloro-
triazine family which also includes simazine, was chosen
* Corresponding author. Tel.:
260061
E-mail address: vessiere@ext.jussieu.fr (A. Vessie`res).
ꢁ
/
33-144-276729; fax:
ꢁ/33-143-
0022-328X/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 3 2 8 X ( 0 2 ) 0 2 0 9 8 - 3

60
N. Fischer-Durand et al. / Journal of Organometallic Chemistry 668 (2003) 59Á66
/
because of its widespread application as a herbicide in
fields of corn and sorghum, due to its low cost and high
efficacy. As a consequence atrazine is frequently de-
tected in ground and surface waters at higher than
authorized levels.
This paper presents our results for the quantification
of atrazine following the CMIA format, exploring the
use of two polyclonal antibodies (anti-atrazine and -
simazine), three cobalt-carbonyl tracers and two differ-
ent buffer solutions.
tion of compounds, a liquid nitrogen cooled 1.0 mm2
InSb (indium antimoine) detector and microbeam (1.0
mm-diameter) for quantitative analysis. FT-IR data
were manipulated on a Windows-equipped PC using
the WinBomemEasy program. Routinely, 44 scans were
coadded in about 1 min and resulting interferogram was
apodized using a cosine function and then Fourier-
transformed to yield a 4 cmꢀ1 resolution spectrum. The
absorbance value was baseline corrected using the
Quant method included in the program. The IR cell
used was an ultralow volume, gold light-pipe cell with a
fill volume of 30 mL and an optical pathlength of 20 mm
[18].
2. Experimental
2.1. General
2.4. Production of rabbit antibodies and purification of
the IgG fraction
Analytical grade atrazine (ATZ) and simazine (SIM)
were purchased from Riedel-de Haen. Cyanuric chlo-
¨
Anti-atrazine and -simazine antibodies were obtained
ride, ethylamine, isopropylamine, o-amino-n-caproic
acid, glycine, 1-ethyl-3-(3-(dimethylaminopropyl)carbo-
diimide (EDAC) and BSA (bovine serum albumin) were
from Sigma. Hydroxysuccinimide and propargylamine
were purchased from Aldrich and dicobalt octacarbonyl
Co2(CO)8 from Strem Chemicals. Solvents were dried
and purified using standard procedures. Flash chroma-
by injecting immunogens BSAÁ
intradermally into New Zealand White rabbits accord-
ing to a previously described procedure [14]. Crude
/
ATZ and BSAÁSIM
/
antisera were stored at ꢀ20 8C and were stable for
/
years. Throughout this study we worked with the
immunoglobulin G (IgG) fractions of antisera which
were obtained by affinity chromatography purification
of the crude sera on Avid Chrom gel (Unisyn Technol-
ogies Inc, Hopkinton, MA, USA) according to the
manufacturer’s instructions.
tography was performed on silica gel 60 (Merck, 40Á63
/
mm). NMR spectra were recorded on BRUKER AC 200
and BRUKER Avance 400 spectrometers. Standard
Greiner flat-bottomed polystyrene microplates were
used in ELISA experiments. Microplates were washed
with BIO-RAD Model 1575 immunowash, and read
with a BIO-RAD Model 550 microplate reader. Melting
points were determined on a Kofler hot stage and are
uncorrected. Mass spectra were recorded on a NER-
MAG R10-10C mass analyser. Microanalyses were
performed by CNRS (ICSN, Gif sur Yvette, France)
or UPMC (SIAR, Paris, France).
2.5. Synthesis of organometallic tracers of atrazine
2.5.1. 2-Chloro-4-(alkylamino)-6-
(carboxymethylamino)-s-triazine (1) and (2): general
procedure
Cyanuric chloride (5 g, 27.1 mmol) was dissolved in
C6H5CH3 (60 ml) and cooled to 0 8C. NaOH (30%, 5
ml) was added, followed by ethylamine (33% in water,
4.3 g, 1.15 equivalents) for 1 or isopropylamine (2.66 ml,
1.15 equivalents) for 2. The mixture was stirred for 15
min at 0 8C, then for 3 h at room temperature (r.t.). This
was followed by dropwise addition of glycine (2.32 g,
1.14 equivalents) dissolved in 5 M NaOH (7 ml), and the
resulting mixture was left to stir overnight at r.t.
Diethylether (100 ml) and water (100 ml) were added.
The aqueous phase was acidified with concd. HCl until
precipitation occurred (pH 3). The precipitate was
filtered off, washed with H2O, and dried. The crude
product was dissolved in Me2SO (75 ml) at 50 8C, then
H2O was added (10 ml) and the solution was allowed to
stand at r.t. until crystallization occurred. The com-
pound was filtered off, washed with H2O and dried to
yield a white powder.
2.2. Buffers
CMIA buffer (pH 8.7) contained glycerol (100 ml),
Na2HPO4×
/
12 H2O (10.7 g), NaCl (9.5 g) and NaN3 (2 g)
per liter of deionized water. CarbonateÁ
/bicarbonate
buffer (coating buffer pH 9.6) contained Na2CO3 (1.4 g)
and NaHCO3 (2.93 g) per liter. Phosphate-buffered
saline (PBS pH 7.4) contained KH2PO4 (0.2 g),
Na2HPO4×
/
12 H2O (1.78 g), NaCl (8 g) and KCl (0.2
phosphate buffer (pH 5.0) con-
12 H2O (7.3 g)
g) per liter. CitrateÁ
/
tained citric acid (5.1 g) and Na2HPO4×
for 500 ml solution.
/
2.3. FT-IR spectroscopy
FT-IR spectra were recorded on a bench-top Bomem
Michelson MB 100 FT spectrometer equipped with a
liquid nitrogen cooled MCT detector for characteriza-
2.5.1.1. 2-Chloro-4-(ethylamino)-6-
(carboxymethylamino)-s-triazine (1). Yield: 2.8 g,
about 44%; m.p.: 215 8C (dec.); H-NMR (200 MHz,
1

N. Fischer-Durand et al. / Journal of Organometallic Chemistry 668 (2003) 59Á
/
66
61
Me2SO): d 1.05 (3H, t, Jꢂ
/
6.5 Hz, CH3 Et); 3.2 (2H, q,
CH2Ã
NH); 8.3 (1H, t, Jꢂ
MHz, Me2SO): d 22.2; 28.2; 40.5; 42.3; 73.0; 81.3; 164.5;
285
[MHꢁ]; Anal. Calc. for C11H15ClN6O: C, 46.73; H,
5.35; N, 29.72. Found: C, 46.96; H, 5.61; N, 29.17%.
Compound 6: Colorless oil which crystallized at 4 8C
(0.17 g, 50%); m.p.: 104 8C; 1H-NMR (250 MHz,
CDCl3): d 1.14 (6H, m, CH3 iPr); 1.31 (2H, m, CH2);
/
CÅ
/
CH); 3.9 (1H, m, HC(CH3)2); 7.7Á
/
7.9 (2H, m,
Jꢂ
/
6.4 Hz, CH2 Et); 3.86 (2H, m, CH2 Gly); 7.9 (2H, m,
NH); 12.5 (1H, broad s, CO2H); 13C-NMR (50 MHz,
Me2SO): d 14.6; 35.45; 42.6; 135.35; 165.3; 165.8; 168;
/
5.5 Hz, NH amide); 13C-NMR (50
165.9; 168.1; 168.9. MS (CIꢁ
/
CH4) m/z: 283Á
/
171.6. MS (CIꢁ
/
NH3) m/z: 232Á
/
234 [MHꢁ].
2.5.1.2. 2-Chloro-4-(isopropylamino)-6-
(carboxymethylamino)-s-triazine (2). Yield: 1.4 g,
about 22%; m.p.: 205 8C (dec.); H-NMR (200 MHz,
1
Me2SO): d 1.08 (6H, d, Jꢂ
s large, CH2 Gly); 3.94 (1H, m, HC(CH3)2); 7.7Á8.0
(2H, m, NH); 12.5 (1H, broad s, COOH); 13C-NMR (50
/
6.5 Hz, CH3 iPr); 3.85 (2H,
1.60 (4H, m, CH2); 2.13 (3H, m, CH2Ã
CH); 3.31 (2H, m, NHÃCH2ÃR); 3.97 (2H, dd, Jꢂ
and 2.5 Hz, CH2ÃCÅCH); 4.08 (1H, m, HC(CH3)2);
5.62 (1H, d, Jꢂ7.75 Hz, NH amine); 6.13 (1H, m, NH
amine); 6.68 (1H, t, Jꢂ
5.87 Hz, NH amide); 13C-NMR
/
CONH and CÅ
/
/
/
/
/
5.25
/
/
MHz, Me2SO): d 22.1; 22.5; 42.4; 164.5; 171.5. MS
248 [MHꢁ].
/
(CIꢁ
/
CH4) m/z: 246Á
/
/
(60 MHz, CDCl3): d 22.7; 23.2; 25.6; 26.8; 29.5; 36.5;
41.0; 43.3; 71.8; 80.1; 165.1; 166.1; 168.3; 173.0. MS
2.5.2. 2-Chloro-4-(isopropylamino)-6-
(carboxypenthylamino)-s-triazine (3)
Previously described [7,19].
(CIꢁ
/
CH4) m/z: 339Á
341 [MHꢁ]; Anal. Calc. for
C15H23ClN6O: C, 53.14; H, 6.84; N, 24.80. Found C,
53.33; H, 6.88; N, 24.52%.
/
2.5.3. 2-Chloro-4-(alkylamino)-6-[N(3-
propynamido)alkylamino]-s-triazines (4), (5), and (6):
general procedure
2.5.4. Complexation of propargylic compounds 4 and 5
with Co2(CO)8: general procedure
The previously prepared acid 1, 2 or 3 (0.907 mmol)
was dissolved in anhydrous DMF (12 ml) under Ar. The
In a three-neck round bottomed flask protected from
light and under argon, a propargylic derivative (0.409
mmol) was dissolved in anhydrous DMF (9 ml).
Co2(CO)8 (0.15 g, one equivalent) was added three times
at 15 min intervals and the reaction was monitored by
resulting solution was cooled to ꢀ
xysuccinimide (0.136 g, 1.3 equivalents), EDAC×
/
15 8C, then hydro-
HCl
/
(0.226 g, 1.3 equivalents), and iPr2EtN (0.205 ml, 1.3
equivalents) were added. The mixture was stirred for 1 h
TLC (EtOAc, Rfꢂ
/
0.8 for 7, Rfꢂ0.9 for 8). After 90
/
at ꢀ
/
15 8C, then for 2 h at 0 8C, and allowed to stand at
min the reaction mixture was directly flash chromato-
graphed under Ar on SiO2 (15 cm length, EtOAc). The
combined fractions containing the expected compound
were concentrated and again flash chromatographed to
yield an orange powder.
r.t. overnight. The colored solution (yellow to orange
depending on the starting acid) was cooled to 0 8C, and
propargylamine (74.7 mL, 1.2 equivalents) was added
under Ar. The solution was stirred at 0 8C for 3 h.
For compound 4, the solution was poured into a flask
cooled on an ice bath, and cold H2O was added (15 ml).
The precipitate was filtered off, washed with cold water
and dried to yield an off-white powder (0.65 g, 65%).
For compounds 5 and 6, the solutions were diluted
with EtOAc (20 ml), and H2O was added (20 ml). The
Compound 7: 50 mg (22%); IR (CHCl3): 2031.6;
1
2058.6; 2097; H-NMR (400 MHz, Me2SO): d 1.0 (3H,
m, CH3); 3.2 (2H, m, CH2 Et); 3.8 (2H, broad s, CH2
Gly); 4.5 (2H, broad d, CH2Ã
/
CÅ
/
CH); 6.6 (1H, s, CÅ
/
CH); 7.8 (2H, m, NH amines); 8.6 (1H, NH amide); 13C-
NMR (100 MHz, Me2SO): d 14.3; 35.0; 40.9; 43.5; 73.0;
95.2; 165.1; 165.8; 167.7; 168.8; 200.0. MS (FABꢁ) m/z:
aqueous phases were extracted with EtOAc (3ꢃ50 ml).
The organic phases were dried (Na2SO4), and evapo-
rated to dryness under reduced pressure. The crude
/
555Á
557 [MHꢁ].
/
Compound 8: 69 mg (30%)ꢁ50 mg of starting
/
products were purified on SiO2 (EtOAc, Rfꢂ
both compounds).
Compound 4: m.p.: 264 8C; 1H-NMR (200 MHz,
Me2SO): d 1.0 (3H, t, Jꢂ7.1 Hz, CH3 Et); 3.1 (1H,
broad s, CÅCH); 3.2 (2H, q, Jꢂ6.7 Hz, CH2 Et); 3.8
/0.6 for
compound; IR (CHCl3): 2031.9; 2058.7; 2097.2; 1H-
NMR (400 MHz, Me2SO): d 1.2 (6H, broad s, CH3
iPr); 4.1 (3H, m, CH2 GlyꢁHC(CH3)2); 4.7 (2H, broad
/
/
s, CH2Ã
/
CÅ
/
CH); 6.4 (1H, broad s, CÅ
/
CH); 6.7Á6.9 (2H,
/
/
/
m, NH amines); 8.1 (1H, broad s, NH amide); 13C-
(4H, m, CH2 Glyꢁ
/
CH2Ã
/
CÅ
/
CH); 7.8Á
/
7.9 (2H, broad m,
NMR (100 MHz, Me2SO): d 22.3; 42.0; 43.1; 44.6; 73.4;
NH amines); 8.32 (1H, broad m, NH amide); 13C-NMR
(50 MHz, Me2SO): d 14.5; 28.2; 35.3; 43.9; 73.0; 81.4;
95.5; 165.7; 167.1; 169.2; 169.5; 200.6; MS (CIꢁ
/
NH3)
m/z: 569Á
/
571 [MHꢁ].
165.2; 166.0; 168.1; 168.9. MS (CIꢁ
/
CH4) m/z: 269Á271
/
[MHꢁ]; Anal. Calc. for C10H13ClN6O: C, 44.69; H,
4.87; N, 31.27. Found: C, 45.05; H, 5.22; N: 30.52%.
Compound 5: Off-white powder (0.1 g, 39%); m.p.:
2.5.5. Complexation of propargylic compound 6 with
Co2(CO)8
In a three-neck round bottomed flask protected from
light and under Ar, propargylic compound 6 (72 mg,
1
213 8C; H-NMR (200 MHz, Me2SO): d 1.07 (6H, m,
CH3 iPr); 3.1 (1H, m, CÅ
/
CH); 3.8 (4H, m, CH2 Glyꢁ
/
2.13ꢃ
10ꢀ4 mol) was dissolved in anhydrous THF (4
/

62
N. Fischer-Durand et al. / Journal of Organometallic Chemistry 668 (2003) 59Á66
/
ml). Co2CO8 (0.12 g, 1.5 equivalents) was added and the
reaction was monitored by TLC (EtOAcÁC5H12: 9/1,
Rfꢂ0.9 for 9). After 1 h, THF was evaporated under
reduced pressure and the crude product was flashed
chromatographed under Ar on SiO2 (15 cm length,
2.6.3. Competitive immunoassays for the different tracers
IgG fraction (125 ml) diluted 1:2500 in PBS-BR was
added to 125 ml of 10 dilutions of ATZ from 0.1 to 1500
ng mlꢀ1 in PBS-BR (thus giving a twofold dilution of
concentrations of both ATZ and antibodies). After
mixing and incubation for 1 h at 4 8C, 100 ml of each
solution was added in duplicate to the wells of the
coated plate. Incubation at r.t. was continued for 2 h,
then the plate was washed four times with PBS-T.
Addition of the HRP-conjugated secondary antibody
and plate reading were performed as described above
(titre determination). Standard curves were obtained by
plotting the B/Bo ratios against the concentration of
ATZ. B was the O.D. at a given concentration and Bo
the O.D. read without ATZ.
/
/
EtOAcÁC5H12: 9/1) to yield compound 9 as a red
/
powder (0.1 g, 80%); IR (CHCl3): 2096; 2058; 2031;
1H-NMR (400 MHz, CD2Cl2): d 1.2 (6H, m, CH3 iPr);
1.4 (2H, m, CH2); 1.6 (2H, m, CH2); 1.7 (2H, m, CH2);
2.2 (2H, t, Jꢂ
NHAr); 4.1 (1H, m, HC(CH3)2); 4.6 (2H, d, Jꢂ
CH2ÃCÅCH); 5.2 (1H, broad s, NH amine); 5.5 (1H,
broad s, NH amine); 5.9 (1H, broad s, NH amide); 6.1
(1H, s, CÅ
CH); 13C-NMR (100 MHz, CD2Cl2): d 21.8;
24.5; 26.1; 28.6; 35.7; 40.4; 41.2; 42.6; 72.1; 93.3; 164.8;
/
7.5 Hz, CH2Ã
/
CONH); 3.4 (2H, m, CH2Ã
/
/
6 Hz,
/
/
/
165.7; 168.0; 171.8; 199.3; MS (CIꢁ
/
NH3) m/z: 625Á627
/
[MHꢁ].
2.6.4. Titration curve by CMIA
500 ml fractions consisting of 30 pmol of tracer 7, 8, or
9 (30 ml of a 10ꢀ6 M solution in EtOH) and eight
dilutions of antibody (from 1:20 to 1:300) in CMIA
buffer were incubated for 2 h at 20 8C. The free fraction
of the tracer was then extracted from the aqueous phase
by addition to each tube of 1 ml CMIA buffer-saturated
isopropylether. The mixtures were vortexed for 30 s,
then 750 ml of each organic phase was transferred to 1.5
ml Eppendorf tube and evaporated to dryness on a
Speed Vac (Savant concentrator). IR analysis were
obtained by dissolving dry residues in 30 ml CCl4. The
FT-IR spectra were immediately recorded on the
spectrometer using the gold light-pipe cell previously
described [18]. The absorbance value of the 2058 cmꢀ1
peak was proportional to the free fraction (F) of the
tracer. The absorbance of the total quantity of tracer (T)
present in the incubation step was obtained by extract-
ing 500 ml fraction containing 30 pmol of tracer without
antibodies. The bound fraction (B) was calculated from
2.6. Immunoassays
2.6.1. Plate coating procedure
An atrazineÁ
was prepared using the way previously described [14] to
coat ELISA plates. bLGÁ
ATZ (100 ml, 1 mg mlꢀ1
solution in freshly prepared NaHCO3ÁNa2CO3 buffer)
/
b-lactoglobulin (bLGÁ
/
ATZ) conjugate
/
/
were added to each well of microtiter plates. The plates
were covered with parafilm and stored at 4 8C over-
night. Plates were then washed three times with PBS
buffer, incubated for 1 h at r.t. with 200 ml of 4%
skimmed milk in PBS buffer to block unoccupied sites
and avoid further non-specific binding of proteins, and
finally washed four times with PBS containing 0.1%
Tween 20 (PBS-T). They can be stored prior use for
several months at 4 8C covered with parafilm.
2.6.2. Titre determination of anti-ATZ and -SIM
antibodies
the difference TÁF. Titration curves were constructed by
plotting the B/T ratio versus the inverse of antibody
dilutions. Titre value was the antibody dilution that
/
Serial dilutions of IgG fractions in PBSꢁ
milk (PBS-BR) were prepared at 100 ml/well and allowed
to react with bLGÁATZ bound to the wells at r.t. for 2
/
2% skimmed
bound 50% of tracer (%B/Tꢂ50).
/
/
h. Plates were then washed four times with PBS-T.
Peroxidase-goat anti-rabbit conjugate at a 1:8000 dilu-
tion in PBS-BR was added at 100 ml/well and allowed to
stand at r.t. for 2 h. Plates were then washed four times
with PBS-T. Substrate consisting of o-phenylenediamine
2.6.5. Standard curve by CMIA (optimized procedure)
500 ml fractions consisting of antibody (IgG fraction)
in PBS buffer at the titre value and 10 amounts of ATZ
(from 12.5 to 1000 pmol) dissolved in 50 ml EtOH were
incubated for 1 h at 4 8C. Then 30 pmol of tracer 9 (30
ml of a 10ꢀ6 M solution in EtOH) was added and
incubation was continued for 2 h at 20 8C. Separation of
the free and bound fractions of the tracer and determi-
nation of the bound fractions were performed as
described for the titration curve. Bo was the value
obtained without atrazine in the incubation step.
Standard curves were then obtained by plotting the B/
Bo ratios versus the amounts of ATZ.
dihydrochloride (OPD, 14 mg) in citrateÁphosphate
/
buffer pH 5.0 (20 ml) with 8 ml H2O2 30%, was added
at 100 ml/well and allowed to stand at r.t. in the dark for
10Á20 min. The reaction was stopped by addition of 50
/
ml/well of H2SO4 2.5 M. After 10 min in the dark, the
plate was read at 492 nm on a microplate reader. 1:5000
dilutions for both anti-ATZ and -SIM IgG fractions,
giving optical densities around 1Á/1.5, were chosen for
the competitive inhibition by ELISA.

N. Fischer-Durand et al. / Journal of Organometallic Chemistry 668 (2003) 59Á
/
66
63
3. Results and discussion
with the same aminocaproic linker used for the pre-
paration of immunogens injected into rabbits for anti-
body production, and therefore 9 would be expected to
be better recognized by antibodies than the two other
tracers which bear a smaller glycine linker. On the other
The CMIA method is a homogenous competitive
immunoassay between the compound to be quantified,
in this case ATZ, and a metal carbonyl tracer, here an
alkyne derivative of atrazine bearing a dicobalt hex-
acarbonyl moiety (Scheme 1). After incubation with
specific antibodies, the free fraction of the tracer is
extracted by isopropylether. After evaporation of the
solvent, the residue is taken in 30 ml CCl4 and quantified
by FT-IR. Quantitative analysis of the tracer is
performed by direct measurement of the absorbance of
the central peak at 2058 cmꢀ1, one of the three
characteristic nCO vibration bands of the metal-carbonyl
unit (Fig. 1) [12,20]: it has been demonstrated that this
hand, a sensitive CMIA assay requires that the tracerÁ
antibody equilibrium constant not be too high, so as not
to overwhelm the atrazineÁantibody equilibrium when
working with trace quantities of atrazine.
Tracers ATZÃC5ÃCo2(CO)6 and ATZÃ
were first evaluated by ELISA technique on microplates
coated with 1 mg mlꢀ1 of bLGÁ
ATZ conjugate. The
/
/
/
/
/
C1Ã
/
Co2(CO)6
/
results of competitive inhibition assays are given in
Table 1. IC50 value is the amount of ATZ or tracer
needed to displace 50% of antibodies from binding to
ATZ coated on plate. The lower this quantity is, the
more specific the antibodies are for the given compound.
Anti-simazine IgG exhibited a higher reactivity towards
ATZ than anti-atrazine IgG: the concentration of free
ATZ required to achieve 50% inhibition was almost
three times lower than with anti-atrazine IgG. Cobalt
carbonyl complexes were recognized by either anti-
atrazine and -simazine IgG. Competitive inhibition
assays carried out with anti-ATZ IgG gave 50% inhibi-
tion at a concentration of 120 pmol mlꢀ1 for tracer 9
which is twofold less reactive than the analyte, but 16.5-
fold more reactive than tracer 8 which behaves as a weak
inhibitor. Experiments with anti-SIM IgG yielded
inhibitory concentrations in the same order of magni-
tude, but the difference in reactivity between the two
tracers was weaker (5.5 rather than 16.5). These results
would suggest that the dicobalt hexacarbonyl entity was
too close to the chlorotriazine moiety in tracer 8 and
therefore weakened the binding affinity of both anti-
ATZ and -SIM antibodies.
peak gives the lowest detection limit. AntibodyÁantigen
/
interactions were performed with polyclonal antibodies
raised against atrazine and simazine immunoconjugates
(Scheme 2) [14]. The two herbicides are too small to be
immunogenic, therefore they were coupled to the lysine
residues of protein, here BSA, by means of a C5-alkyl
chain linker. ELISA and CMIA analyses were per-
formed with the IgG fraction of sera obtained by
purification on affinity chomatography of crude sera.
Our previous results [15] had shown that atrazine was
well recognized by both anti-ATZ and -SIM antibodies.
In a preliminary study, rabbit antisera collected every 2
weeks were screened by ELISA on plates coated with
bLGÁ
/
ATZ in order to select sera exhibiting the highest
concentrations of specific anti-ATZ antibodies.
The three dicobalt hexacarbonyl tracers were pre-
pared following the general strategy depicted in Scheme
3. Cyanuric chloride was first reacted with ethylamine
(for the simazine tracer) or isopropylamine (for the
atrazine tracer) at 0Á20 8C, then either o-amino-n-
/
caproic acid or glycine was added at 20 8C, leading to
acid derivatives 1, 2, 3. Activation of acid moieties with
hydroxysuccinimide, then nucleophilic addition of pro-
pargylamine at 0 8C yielded the three propargylic
derivatives 4, 5, 6, which were then complexed with
dicobalt octacarbonyl to give the three expected tracers
In another set of experiments, tracers 7, 8, and 9 were
evaluated in the determination of the titre value of
antibodies to be used in the competitive assay by CMIA.
A titration curve by CMIA was obtained after incubat-
ing 30 pmol of organometallic tracer with decreasing
amounts of antibodies in CMIA buffer. The amount of
free tracer extracted by isopropylether increased as the
7, 8, 9. Tracer 9, ATZÃ
/
C5Ã
/
Co2(CO)6, was prepared
Scheme 1. Principle of the CMIA method.

64
N. Fischer-Durand et al. / Journal of Organometallic Chemistry 668 (2003) 59Á66
/
Fig. 1. FT-IR spectrum of the dicobalt hexacarbonyl complex of ATZ (9). 30 pmol of tracerꢁ
ml iPr2O. 750 ml were transferred to 1.5 ml Eppendorf microtube and evaporated to dryness, then taken in 30 ml CCl4 for FT-IR quantification.
/
50 ml EtOHꢁ420 ml CMIA buffer. Extraction with 1
/
Table 1
a
IC50
IC50 (pmol mlꢀ1
)
b
b
Anti-ATZ antibody
65
Anti-SIM antibody
ATZ
ATZÃ
ATZÃ
25
200
/
C5Ã
/
(Co)2(CO)6 9 120
/
C1Ã
/(Co)2(CO)6 8 2000
1100
Scheme 2. Chemical structures of atrazine, simazine and immunocon-
jugates.
a
Concentration of ATZ, tracers 9 and 8 required to displace 50% of
antibodies from binding to ATZ coated on plate.
b
IgG fraction.
quantity of antibodies decreased. As a result, the height
of the peak at 2058 cmꢀ1 increased proportionally to
the free fraction of the tracer. The titre is defined as the
antibody dilution that binds 50% of the tracer. The
highest the dilution is, the more specific the antibodies
both anti-ATZ and -SIM antibodies to reach 50%
binding, at dilutions 100 for anti-ATZ and 74 for anti-
SIM. In the case of tracers with short spacer, even with
high concentration of antibodies (i.e. dilution 20), we
did not reach 50% binding. Moreover, we could not
perform the assay at higher antibody concentration
without encountering strong emulsions, preventing
are for binding to a given tracer. Only tracer ATZÃ
/
C5Ã
/
Co2(CO)6 exhibited sufficient binding affinity towards
Scheme 3. General synthesis of tracers.

N. Fischer-Durand et al. / Journal of Organometallic Chemistry 668 (2003) 59Á
/
66
65
Fig. 2. Competition curves by CMIA, CMIA buffer pH 8.7, without
pre-incubation: (A) anti-ATZ IgG fraction, dilution 1:100; (B) anti-
SIM IgG fraction, dilution 1:74.
Fig. 3. Competition curves by CMIA: (A) anti-ATZ IgG fraction, pre-
incubation for 1 h at 4 8C, CMIA buffer ph 8.7; (B) anti-SIM IgG
fraction, pre-incubation for 1 h at 4 8C, PBS buffer pH 7.4.
proper extraction of the tracer. This drop in binding
affinity observed with the glycine spacer was consistent
with the first results obtained by the ELISA technique
(see above), and proved the need for at least a medium-
length spacer arm between the triazine ring and the
mined in a 500 mL solution assay were incubated for 2 h
at room temperature. Extraction of the free tracer and
quantification by FT-IR gave the standard curves
shown in Fig. 2. The competition curves obtained for
anti-ATZ antibodies showed that ATZ and tracer 9 did
not efficiently compete for binding to the antibodies
when added at the same time, while in the case of anti-
SIM antibodies, the two compounds competed well in
metal carbonyl moiety in order to retain tracerÁ
body affinity.
Accordingly, calibration curves for atrazine quantita-
tion were conducted with ATZÃC5ÃCo2(CO)6.
/anti-
/
/
the range 0Á200 pmol/tube of ATZ, but at higher
/
In previous CMIA for antiepileptic drugs [11,12], the
incubation step was carried out in CMIA buffer pH 8.7
without pre-incubation of the analyte with antibodies.
Our first experiments were performed under these
concentrations, ATZ did not displace tracer 9, suggest-
ing that the antibodies recognized the spacer arm
common to immunogen BSAÁATZ and tracer 9.
/
In light of these results, we decided to perform a pre-
incubation step, which is commonly used in immunoas-
says [21,22]. The assays were run with anti-ATZ
conditions. Tracer (30 pmol), ATZ (12.5Á
/500 pmol)
and antibodies at the titre dilution previously deter-

66
N. Fischer-Durand et al. / Journal of Organometallic Chemistry 668 (2003) 59Á66
/
antibodies in CMIA buffer pH 8.7. IgG fraction at 1:100
dilution was incubated for 1 h at 4 8C with 12.5Á500
References
/
[1] S. Lacorte, I. Guiffard, D. Fraisse, D. Barcelo, Anal. Chem. 72
(2000) 1430.
pmol/tube of ATZ. Then 30 pmol of tracer 9 was added,
and incubation continued for 2 h at room temperature
before extraction of the free tracer. Standard curves
plotted in Fig. 3A showed an improvement in the shape
of the curve, though high unspecific interactions still
existed (about 40%).
Finally, changing from CMIA buffer pH 8.7 to PBS
buffer pH 7.4, and using anti-SIM antibodies led to
reproducible standard curves (Fig. 3B). IC50 was about
200 pmol/tube (86 mg lꢀ1) and the sensitivity (IC80) was
about 50 pmol/tube (21 mg lꢀ1). These values are far
from the sensitivity required by European regulations to
reach 0.1 mg lꢀ1 but our results were obtained without
the help of any sample pre-concentration which is
routinely used with HPLC or GC analyses. On the
other hand, an interesting feature of this immunoassay
compared to standard ones is its capacity for simulta-
neous multiple assays as previously described [11], when
using tracers with different metal carbonyl moieties.
[2] R. Curini, A. Gentili, S. Marchese, A. Marino, D. Perret, J.
Chromatogr. A 874 (2000) 187.
[3] E. Hogendoorn, P. Van Zoonen, J. Chromatogr. A 892 (2000)
435.
[4] D. Stevenson, J. Chromatogr. B 745 (2000) 39.
[5] D.H. Thomas, M. Beck-Westermeyer, D.S. Hage, Anal. Chem. 66
(1994) 3823.
[6] J.M. Van Emon, C.L. Gerlach, K. Bowman, J. Chromatogr. B
715 (1998) 211.
[7] B. Dunbar, B. Riggle, G. Niswender, J. Agric. Food Chem. 38
(1990) 433.
[8] J.M. Van Emon, V. Lopez-Avila, Anal. Chem. 64 (1992)
79A.
[9] V. Philomin, A. Vessie`res, G. Jaouen, J. Immunol. Methods 171
(1994) 201.
[10] M. Salmain, A. Vessie`res, P. Brossier, I.S. Butler, G. Jaouen, J.
Immunol. Methods 148 (1992) 65.
[11] M. Salmain, A. Vessie`res, A. Varenne, P. Brossier, G. Jaouen, J.
Organomet. Chem. 589 (1999) 92.
[12] A. Varenne, A. Vessie`res, M. Salmain, P. Brossier, G. Jaouen, J.
Immunol. Methods 186 (1995) 195.
[13] J.F. Lawrence, C. Menard, M.-C. Hemion, V. Pichon, F. Le
Goffic, N. Durand, J. Chromatogr. A 752 (1996) 147.
[14] V. Pichon, L. Chen, M.C. Hennion, R. Daniel, A. Martel, F. Le
Goffic, J. Abian, D. Barcelo, Anal. Chem. 67 (1995)
2451.
4. Conclusion
In this study, we have shown the feasibility of using
the non-isotopic CMIA method in environmental mon-
itoring, by modification of the procedure already
developed for the clinical diagnostics. Pre-incubation
in PBS buffer pH 7.4 of anti-SIM antibodies with
increasing amounts of atrazine to be quantified for 1 h
at 4 8C, followed by addition of 30 pmol of dicobalt
hexacarbonyl tracer 9 and incubation for 2 h at room
temperature, led to optimized standard curves as shown
in Fig. 3B. At present the carbonyl metallo immunoas-
say is not sensitive enough to satisfy the European
Guidelines, but our first results are encouraging, bearing
in mind that this assay is carried out without sample pre-
treatment.
[15] V. Pichon, L. Chen, N. Durand, F. Le Goffic, M.-C. Hennion, J.
Chromatogr. A 725 (1996) 107.
[16] V. Pichon, H. Rogniaux, N. Fischer-Durand, S. Ben Rejeb, F. Le
Goffic, M.-C. Hennion, Chromatographia 45 (1997) 289.
[17] A. Vessie`res, F. Le Bideau, N. Fischer-Durand, P. Janvier, J.-M.
Heldt, S. Ben Rejeb, G. Jaouen, Appl. Organometal. Chem. 16
(2002) (2002) 669Á674.
/
[18] A. Vessie`res, G. Jaouen, M. Salmain, I.S. Butler, Appl. Spectrosc.
44 (1991) 1092.
[19] B. Dunbar, G. Niswender, J. Hudson, U.S. Patent 4,530,786,
1985.
[20] M. Salmain, A. Vessie`res, G. Jaouen, I.S. Butler, Anal. Chem. 63
(1991) 2323.
[21] W.H. Newsome, J.M. Yeung, P.G. Collins, J. AOAC Int. 76
(1993) 381.
[22] S. Ben Rejeb, N. Fischer-Durand, A. Martel, F. Le Goffic, J.F.
Lawrence, J.M. Yeung, M.A. Abbott, Int. J. Environ. Anal.
Chem. 69 (1998) 13.
Increasing the sensitivity by amplification of the FT-
IR signal is under investigation and will be reported in
due course.