Use of L-lysine fluorescence derivatives as tracers to enhance the performance of polarization fluoroimmunoassays. A study using two herbicides as model antigens

Article abstract of DOI:10.1021/ac011051x

Fluorescence polarization immunoassay (FPIA) is a convenient homogeneous assay, the use of which is restricted in environmental analysis by low sensitivity and matrix effects. We selected the herbicides 2,4D and 2,4,5T to synthesize new L-lysine-based flu

Full text of DOI:10.1021/ac011051x

Anal. Chem. 2002, 74, 2513-2521
Use of L-Lysine Fluorescence Derivatives as
Tracers To Enhance the Performance of
Polarization Fluoroimmunoassays. A Study Using
Two Herbicides as Model Antigens
George I. Hatzidakis,*,† Aristidis M. Tsatsakis,*,† Elias K. Krambovitis,‡ Apostolos Spyros,§ and
|
Sergei A. Eremin
Laboratory of Toxicology, Medical School, University of Crete, Voutes, Heraklion 71110, Crete, Greece, Department of
Applied Biochemistry and Immunology, Institute of Molecular Biology and Biotechnology, P.O. Box 1527, Heraklion 71110,
Crete, Greece, Chemistry Department, Lomonosov Moscow State University, 119899 Moscow, Russia, and
NMR Laboratory, Department of Chemistry, University of Crete, Heraklion 71409, P.O Box 1470, Crete, Greece
or for screening purposes.7 Among these, the homogeneous
methods are fast, simple, and inexpensive and they do not require
separation or washing steps.
Fluorescence polarization immunoassay (FP IA) is a con-
venient homogeneous assay, the use of which is restricted
in environmental analysis by low sensitivity and matrix
effects. We selected the herbicides 2 ,4 D and 2 ,4 ,5 T to
The fluorescence polarization assay (FPIA) is the most widely
used homogeneous assay with applications mainly in therapeutic
and drug abuse monitoring.8-10 In recent years a lot of work has
been performed for FPIA development and application in envi-
ronmental analysis. Pesticides such as 2,4D, 2,4,5T, atrazine,
simazine, isoproturone, methabenzthiazurone, dichlorprop, and
propazine have been used as antigens.6,8,11,12 Despite the apparent
advantages, the relatively low sensitivity remains as the major
limitation of the assay.8 This is particularly important for residue
analysis where the limits of detection (LODs) are very low and
the sample matrixes are complex. Efforts for further FPIA
optimization are directed toward improvement in sensitivity with
the use of new more suitable materials and fluorescent dyes,13,14
new instruments,8 and new and alternative methodologies such
as 2-PEFA (2-photon excited fluorescence anisotropy)15 or stopped-
flow FPIA,16,17 where the matrix effect is minimized.
synthesize new L-lysine-based fluorescent tracers using
solid-phase chemistry. In addition, three different immu-
nogens of 2 ,4 ,5 T were prepared for immunization and
antibody production. The new tracers and antibodies were
adapted to FP IA. Tracers with the hapten attached to the
r-aminogroup of L-lysine and fluorescein to the e-amino
group exhibited at least a 5-fold increased sensitivity when
compared to the previously reported ethylenediamine-
based tracer (2 ,4 D-EDA-F). The isomeric structure (hap-
ten attached to the e-amino and fluorescein to the a-amino
group) appeared 7 .6 times less sensitive, and all other
alternative structures exhibited even lower sensitivities.
This observation was confirmed against the monoclonal
anti-2 ,4 D antibody E2 / G2 and polyclonal anti-2 ,4 ,5 T
antibodies. The affinity constant of 2 ,4 D-EDA-F with E2 /
G2 was 8 .1 times higher when compared with the new
(2) Cuong, N. V.; Bachman, T. T.; Schmid, R. D. Fresenius J. Anal. Chem. 1 9 9 9 ,
364, 584-9.
(3) Yazinina, E. V.; Zherdev, A. V.; Dzantiev, B. B.; Izumrudov, V. A.; Gee, S. J.;
Hammock, B. D. Anal. Chem. 1 9 9 9 , 71, 3538-43.
(4) Bjarnason, B.; Bousios, N.; Eremin, S.; Johansson, G. Anal. Chim. Acta
1 9 9 7 , 347 (1-2), 111-20.
tracer, suggesting the more specific nature of the L-lysine-
based tracer, the use of which leads to a more sensitive
assay. This type of tracer could improve performance and
lower substantially the detection limits of FP IAs.
(5) Gonzalez-Martinez, M. A.; Morais, S.; Puchades, R.; Maquieira, A.; Abad,
A.; Montoya, A. Anal. Chem. 1 9 9 7 , 69 (14), 2812-18.
(6) Onnerfjord, P.; Eremin, S.; Emneus, J.; Marko-Varga, G. J. Immunol. Methods
1 9 9 8 , 213, 31-9.
(7) Nistor, C.; Emneus, J. Waste Manage. 1 9 9 9 , 19, 147-70.
(8) Eremin, S. A. Food Technol. Biotechnol. 1 9 9 8 , 36 (3), 235-43.
(9) Schwenzer, K. S.; Pearlman, R.; Tsilimidos, M.; Salamone, S. J.; Canon, R.
C.; Wong, S. H.; Gock, S. B.; Jentzen, J. J. J. Anal. Toxicol. 2 0 0 0 , 24 (8),
726-32.
(10) Eremin, S. A.; Zaitchev, S. V.; Egorov, A. M.; Ermakov, A. N.; Smirnov, A.
V.; Izotov, B. N. Anal. Chim. Acta 1 9 8 9 , 227, 287-90.
(11) Melnichenko, O. A.; Eremin, S. A.; Egorov, A. M. J. Anal. Chem. 1 9 9 6 , 51
(5), 512-5.
A large number of heterogeneous and homogeneous immu-
noassays have been developed for detection of biologically active
residual compounds responsible for food and environmental
contamination. They include ELISA with colorimetric, chemilu-
minescent, or electrochemical detection, biosensors and immu-
nosensors, flow-injection immunoassays, and fluorescence im-
munoassays.1-6 These are used for direct detection of the analyte
* Corresponding authors. Tel: +30-81-394679. Fax: +30-81-542098. E-mail:
aris@med.uoc.gr (A.M.T.); ghatzi@imbb.forth.gr (G.I.H.).
† Medical School, University of Crete.
‡ Institute of Molecular Biology and Biotechnology.
§ Department of Chemistry, University of Crete.
| Lomonosov Moscow State University.
(12) Schobel, U.; Barzen, C.; Gauglitz, G. Fresenius’ J. Anal. Chem. 2 0 0 0 , 366
(6-7), 646-58.
(13) Sun, W.-C.; Gee, K. R.; Klaubert, H.; Haugland, R. P. J. Org. Chem. 1 9 9 7 ,
62, 6469-75.
(1) Dzgoev, A. B.; Gazaryan, I. G.; Lagrimini, L. M.; Ramanathan, K.; Danielsson,
B. Anal. Chem. 1 9 9 9 , 71 (22), 5258-61.
(14) Aguilar-Caballos, M. P.; Gomez.-Henz, A.; Perez-Bendito, D. Anal. Chim.
Acta 1 9 9 9 , 381, 147-54.
10.1021/ac011051x CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/27/2002
Analytical Chemistry, Vol. 74, No. 11, June 1, 2002 2513

The specific antibody used is one of the key components of
FPIA, like in any other immunoassay, that determines the quality
of the method.18-20 In FPIA, however, the tracer structure also
greatly affects the sensitivity. Tracers of 2,4D with different bridge
lengths (2, 4, or 6 carbon atoms) have been compared using the
same antibody. The most sensitive assay was obtained using the
ethylenediamine (2 carbon atom chemical bridge) based tracer.8,21
Similar findings have been reported for atrazine FPIA.6,22 Therefore
these kinds of tracers have been used for FPIA optimization.16,17,23
In this work, we used the herbicides (2,4-dichlorophenoxy)-
acetic acid (2,4D) and (2,4,5-trichlorophenoxy)acetic acid (2,4,-
5T) as model analytes and studied the influence of tracer structure
in FPIA characteristics, mainly sensitivity. 2,4D is a widely used
herbicide that could potentially contaminate the ground and water
and enter the food chain. The presence of dioxin traces in 2,4,5T
has led to its withdrawal from the market and its use forbidden.
The importance of this molecule remains high because monitoring
of 2,4,5T levels serves as a potential indicator of dioxin pollution.24
Dibromophenoxy)acetic acid (2,4DBr) and other structurally
related compounds for cross-reaction studies were provided from
Riedel-de-Haen (Hannover, Germany). Monoclonal anti 2,4D
antibodies (E2/ G2) were obtained at the VRI (Brno, Czech
republic) as elsewhere described25 and kindly provided by Dr.
Milan Franek. All reagents were of analytical grade. The water in
all experiments was double distilled water, and the buffer for FPIA
was the TDx buffer of Abbot Diagnostics. Stock solutions of
analytes and cross-reactants were prepared in methanol. TLC for
analysis and purification of products and ninhydrin tests was
executed as elsewhere described.27 The TLC solvent system in
all cases was 7:1.5:1.5 toluol-methanol-acetic acid.
Abbreviations: Fmoc, fluorenylmethyloxycarbonyl; Boc, butyl-
oxycarbonyl; Mtt, methoxytrityl; âAla, â-alanine; ꢀAca, ꢀ-amino-
caproic acid; K,
Synthetic P rocedures. The synthetic process we followed
included (a) synthesis of -lysine and ethylenediamine (EDA)
L-lysine.
L
derivatives of the analytes, using 2-chlorotrityl chloride solid-phase
chemistry, (b) labeling of all derivatives with FITC for tracers
preparation, and (c) synthesis of 2,4,5T haptens with solid-phase
chemistry and preparation of immunogens.
We introduced L-lysine as a spacer for preparation of monolabeled
and bilabeled fluorescent tracers of 2,4D and 2,4,5T by applying
solid-phase chemistry. In addition, three different bridge length
2,4,5T immunogens were synthesized. The monoclonal antibodies
clone E2/ G2,25 specific for 2,4D and all polyclonal antibodies
produced against 2,4,5T immunogens, were adapted to FPIA in
combination with the new tracers. The corresponding ethylene-
diamine-based tracers were also used for comparison purposes.
The results of these assays are presented in this study, indicating
One Lysine Molecule Resin, H2N-K(Boc)-CLTR (1), and Two
Lysine Molecule Resin, H2N-K2(Boc)2-CLTR (2). The very acidic
labile CLTR resin was used as the solid phase. Synthesis started
with 1 g of CLTR resin and 280 mg (0.6 mmol) of Fmoc-K(Boc)-
OH, and the well-established protocol for CLTR loading was
followed.28 The yielded resin had a substitution of 0.4 mmol of
Fmoc-K(Boc)-OH/ g of CLTR. It was then treated with 25%
piperidine solution for 30 min to remove the Fmoc group and
washed successively with dimethyl formamide (DMF), 2-propanol,
and ether. The resulting resin H2N-K(Boc)-CLTR (1 ) was dried
in a nitrogen stream.
that a new particular L-lysine-based tracer structure is important
in improving further the FPIA sensitivity.
EXPERIMENTAL SECTION
Materials. 2-Chlorotrityl chloride resin (CLTR) (1.5 mmol of
reactive Cl/ g of resin) and all Fmoc-protected amino acids, Fmoc-
K(Boc)-OH, Fmoc-K(Fmoc)-OH, Fmoc-K(Mtt)-OH, Fmoc-âAlR-
OH, and Fmoc-ꢀAca-OH, were purchased from CBL, Patras,
Greece. Dicyclohexyl carbodiimide (DCC), hydroxybenzotriazole
(HOBt), N-hydroxysuccinimide (NHS), and all other chemicals
and organic solvents were purchased from Merck (Darmstadt,
Germany). (2,4-Dichlorophenoxy)acetic acid (2,4D) (UVmax ) 284
nm, ꢀ1M ) 2100) and (2,4,5-trichlorophenoxy)acetic acid (2,4,5T)
(UVmax ) 289 nm, ꢀ1M ) 2550)26 of analytical grade and fluorescein
5-isothiocyanate isomer I (FITC) were purchased from Sigma. (2,4-
A 300 mg amount of resin 1 used for preparation of resin 2 .
A 170 mg (0.36 mmol) amount of Fmoc-K(Boc)-OH in 3 mL of
DMF and DCC/ HOBt (82 mg/ 82 mg) as the condensing agent
was used28 in a 2 h reaction with resin 1 . The completion of the
reaction was checked with TLC.27 Resin wash and Fmoc removal
as above were followed. The resulting resin H2N-K2(Boc)2-CLTR
(2 ) was dried in a nitrogen stream.
2,4D-K(NH2)-OH (3a ), 2,4,5T-K(NH2)-OH (3b), 2,4D-K2-
(NH2)2-OH (4a ), and 2,4,5T-K2(NH2)2-OH (4b). A 100 mg
amount of each resin 1 and 2 with approximately 0.04 mmol of
amino groups each was used for reaction with 0.2 mmol of 2,4D
and 2,4,5T. The condensing agent NHS/ DCC (26 mg/ 46 mg) was
used for a 1 h reaction of herbicides with resins 1 and 2 . After
the usual washes, resins were treated for 30 min with 1 mL of
90% TFA to cleave haptens and simultaneous Boc deprotection.
The filtrates evaporated in nitrogen stream and the products 2,4D-
K(NH2)-OH (3 a), 2,4,5T-K(NH2)-OH (3 b), 2,4D-K2(NH2)2-OH
(4a), and 2,4,5T-K2(NH2)2-OH (4b) (Figure 1) were easily isolated
with diethyl ether precipitation, centrifugation, and drying. The
ninhydrin test on TLC plates in all cases was positive, Rf3 a ) 0.2,
Rf3 b ) 0.25, Rf4 a ) 0.01, and Rf4 b ) 0.02. UVmax of 3 a and 4 a
(15) Baker, G. A.; Pandey, S.; Bright, F. V. Anal. Chem. 2 0 0 0 , 72 (22), 5748-
52.
(16) Eremin, S. A.; Matveeva, E. G.; Gomez-Hens, A.; Perez-Bendito, D. Int. J.
Environ Anal. Chem. 1 9 9 8 , 71 (2), 137-46.
(17) Sendra, B.; Panadero, S.; Eremin, S.; Gomez-Hens, A. Talanta 1 9 9 8 , 47,
153-60.
(18) Franek, M.; Pouzar, V.; Kolar, V. Anal. Chim. Acta 1 9 9 7 , 347, 163-76.
(19) Hatzidakis, G.; Stefanakis, A.; Krambovitis, E. J. Reprod. Fertil. 1 9 9 3 , 97,
557-61.
(20) Li, K.; Chen, R.; Zhao, B.; Liu, M.; Karu, A. E.; Roberts, V. A.; Li, Q. X.
Anal. Chem. 1 9 9 9 , 71, 302-9.
(21) Eremin, S. A. Polarization fluoroimmunoassays for rapid, specific detection
of pesticides. In Immunoanalysis of Agrochemicals: Emerging Technologies;
Nelson, J. O., Karu, A. E., Wong, R., Eds.; ACS Symposium Series 586;
American Chemical Society: Washington, DC, 1995; pp 223-34.
(22) Eremin, S. A.; Samsonova, J. V. Anal. Lett. 1 9 9 4 , 27, 3013-17.
(23) Matveeva, E. G.; Melik-Nubarov, N. S.; Miethe, P.; Levashov, A. V. Anal.
Biochem. 1 9 9 6 , 234, 13-8.
(26) Moffat, A. C.; Jackson, J. V.; Moss, M. S.; Widdop, B. Clarke’s Isolation and
Identification of Drugs, 2nd ed.; London, 1986.
(24) Medyantseva, E. P.; Vertlib, M. G.; Kutyreva, M. P.; Khaldeeva, E. I.;
Budnikov, G. K.; Eremin, S. A. Anal. Chim. Acta 1 9 9 7 , 347 (1-2), 71-6.
(25) Franek, M.; Kolar, V.; Granatova, M.; Nevorankova, Z. J. Agric. Food Chem.
1 9 9 4 , 42, 1369-71.
(27) Barlos, K.; Gatos, D.; Kapolos, S.; Poulos, C.; Schafer, W.; Wenqing, Y. Int.
J. Pept. Protein Res. 1 9 9 1 , 38, 555-61.
(28) Krambovitis, E.; Hatzidakis, G.; Barlos, K. J. Biol. Chem. 1 9 9 8 , 273 (18),
10874-9.
2514 Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

Figure 1. Chemical structures of synthesized derivatives, tracers, and immunogens. For each compound a code number (in parentheses)
and a short name (with boldface letters) are given. The short name has been chosen to describe the corresponding chemical structure. When
an R group is noted as 2,4D or 2,4,5T, it represents their acyl derivatives. When R is given as FITC, it represents the thiocarbamoyl fluorescein.
1
The notation with Greek characters in L-lysine and with Latin characters in 2,4D and FITC is used for the study of the H NMR chemical shifts
and for the study of the Figures S1-S4 of the Supporting Information. Replacing the two chlorine atoms from 15a with two bromine atoms can
confirm the structure of tracer 15c.
was 284 nm, and the UV spectrum was identical to that of 2,4D,
UVmax of 3 b and 4 b was 289 nm, and UV spectrum was identical
to that of 2,4,5T. Derivatives 3 a and 4 a were further characterized
with NMR spectroscopy. 3a 1H NMR (MeOH-d4): δ ) 1.39 (m,
2H, CH2-γ), 1.65 (m, 2H, CH2-δ), 1.83 (td, 2H, CH2-â), 2.87 (t, 2H,
CH2-ꢀ), 4.35 (t, 1H, CH-R), 4.59 (s, 2H, OCH2CO), 7.05 (d, 1H,
H-y), 7.25 (dd, 1H, H-z), 7.42 (d, 1H H-x). 4a 1H NMR (MeOH-
d4): δ ) 1.44 (m, 4H, CH2-γ,γ′), 1.64 (m, 4H, CH2-δ,δ′), 1.75 (td,
2H, CH2-â), 1.81 (td, 2H, CH2-â′), 2.86 (m, 4H, CH2-ꢀ,ꢀ′), 4.28 (m,
1H, CH-R), 4.43 (t, 1H, CH-R′) 4.61 (s, 2H, OCH2CO), 7.02 (d,
1H, H-y), 7.25 (dd, 1H, H-z), 7.42 (d, 1H H-x). The notation of the
protons is given in Figure 1.
2,4D-EDA-NH2 (5a ), 2,4,5T-EDA-NH2 (5b), and 2,4DBr-EDA-
NH2 (5c). A 10 mL volume of DMF and 1 mL of EDA were added
in 1 g of CLTR under agitation and left 30 min for EDA
immobilization on resin. The resulting resin (H2N-EDA-CLTR)
washed as usually and dried. A 0.45 mmol amount of each analyte
(2,4D, 2,4,5T, and 2,4DBr) was used for reaction with 100 mg of
H2N-EDA-CLTR and the condensing agent NHS/ DCC. Haptens
2,4D-EDA-NH2 (5a), 2,4,5T-EDA-NH2 (5b) (Figure 1), and 2,4DBr-
EDA-NH2 (5 c) were isolated as above. All derivatives exhibited a
positive ninhydrin test with Rf5 a ) 0.39 and Rf5 b ) 0.42. The
UV spectra of 5 a,b were identical to those of 2,4D and 2,4,5T,
respectively.
H2N-K(Mtt)-EDA-2,4D (6a ) and H2N-K(Mtt)-EDA-2,4,5T (6b).
Equal molar quantities of Fmoc-K(Mtt)-OH and the derivatives
5 a,b were used for a 24 h reaction using DCC/ HOBt as the
condensing agent. The products were treated for 30 min with 25%
isopropylamine in dichloromethane (DCM) for the Fmoc removal,
and the derivatives H2N-K(Mtt)-EDA-2,4D (6 a) and H2N-K(Mtt)-
EDA-2,4,5T (6b) were isolated with diethyl ether precipitation and
preparative TLC. Spots with Rf6 a ) 0.43 and Rf6 b ) 0.47 which
give a positive ninhydrin test were collected and products eluted
with methanol.
Analytical Chemistry, Vol. 74, No. 11, June 1, 2002 2515

H2N-K(NH2)-EDA-2,4D (7a ) and H2N-K(NH2)-EDA-2,4,5T
(7b). Derivatives 7 a,b were easily prepared as described in the
previous paragraph using the L-lysine derivative Fmoc-K(Fmoc)-
Table 1. Results from 2,4D FPIA, Obtained Using All
Synthesized Tracers and the E2/G2 Monoclonal
Antibody: Comparison of Sensitivities
OH instead of Fmoc-K(Mtt)-OH. With this modification two amino
groups were incorporated in the same molecule. Spots with
Rf7 a ) 0.04 and Rf7 b ) 0.06 were collected from preparative TLC.
UV spectra of 7 a,b were identical to those of 2,4D and 2,4,5T,
respectively.
concn
(nM)a
I50
tracer
(ng/ mL)b
ratioc
2,4DK(F)-OH (1 3 a)
2,4D-EDA-F (1 5 a)
2
2
2
2
2
8
4
12.2
65
153.8
93.5
111.9
87.8
102.7
1
5.3
12.6
7.6
9.17
7.2
8.4
2,4DBr-EDA-F (1 5 c)
FK(2,4D)-OH (1 2 a)
FK(NH2)-EDA-2,4D (1 6 a)
2,4DK2(F)2OH (1 4 a)
FK(F)-EDA-2,4D (1 7 a)
F-K(2,4D)-OH (12a ) and F-K(2,4,5T)-OH (12b). The pro-
tected L-lysine derivative Fmoc-K(Mtt)-OH (0.3 mmol) was loaded
on 500 mg of CLTR as described above. Fmoc was removed, and
the resulting resin washed and dried as usually. The free R-amino
group reacted with FITC. Briefly, 4.7 mg of FITC dissolved in 0.5
mL of DMF was mixed with 20 mg of resin (H2N-K(Mtt)-CLTR)
and allowed for react for 1 h . The usual washes followed, and
the fluorescent product was cleaved from resin and Mtt depro-
tected, with 1% TFA solution in DCM for 20 min. The filtrate
evaporated, and the fluorescent derivative F-K(NH2)-OH was
isolated with ether precipitation, ether wash, and drying (3.5 mg).
A 0.0065 mmol amount from each 2,4D and 2,4,5T was activated
with the condensing agent NHS/ DCC (1 mg/ 1.5 mg) in 70 µL of
DMF and reacted in two separate reaction vessels with 0.004 mmol
of the F-K(NH2)-OH derivative for 24 h. DMF was evaporated
under vacuum, and 2 mL of diethyl ether was added in each tube
for precipitation of tracers. The tracers F-K(2,4D)-OH (1 2 a) and
F-K(2,4,5T)-OH (1 2 b) (Figure 1) were washed with ether and
dried. They were purified with preparative TLC (Rf1 2 a ) 0.4 and
Rf1 2 b ) 0.42) and stored in methanolic solution.
a
b
Final concentration in reaction mixture. I50: concentration of 2,4D
that reduces the initial polarization (mP0) at 50%. c Ratio: I50 of a tracer/
I50 of 1 3 a. This ratio is an index of the relative sensitivity between
tracers.
(s, 2H, H-n). The structures of all tracers and notations of the
protons are given in Figure 1.
2,4,5T Haptens 2,4,5T-âAla-OH (20) and 2,4,5T-ꢀAca-OH
(21). Two portions of 200 mg of CLTR each were loaded with
Fmoc-âAla-OH (0.2 mmol) and Fmoc-ꢀAca-OH (0.2 mmol) and
treated as described above. Resins H2N-âAla-CLTR and H2N-ꢀAca-
CLTR had a substitution of ∼0.5 mmol/ g. Herbicide 2,4,5T reacted
in a 5 molar excess with both resins as already described. Haptens
were isolated after cleavage with 90% TFA in water. The TFA was
evaporated, and the products were precipitated with water, washed
two times, and dried under vacuum. Derivatives 2 0 and 2 1
(Figure 1) exhibited UVmax ) 289 nm and spectra identical to that
of 2,4,5T.
Preparation of Fluorescent Tracers 13a ,b, 14a ,b, 15a -c,
16a ,b, and 17a ,b. The derivatives 3 a,b, 4 a,b, 5 a-c, 6 a,b, and
7a,b were dissolved in MeOH and reacted with FITC.29 The molar
ratio used in reaction was FITC:amino groups ) 1.5:1. For 0.06
mmol of amino groups, 0.5 mL of MeOH and 5 µL of diisopropyl
ethylamine (DIPEA) were used. After the completion of reaction,
solvent was evaporated and tracers were easily precipitated with
diethyl ether. The precipitates were washed with ether and
purified with preparative TLC. In the case of the 6 a,b derivatives,
the Mtt moiety was removed just after the reaction with FITC,
with 1% TFA in DCM for 20 min. Rf1 3 a ) 0.35, Rf1 3 b ) 0.38,
Rf1 5 a ) 0.52, Rf1 5 b ) 0.53, Rf1 6 a ) 0.15, Rf1 6 b ) 0.16,
Rf1 4 a ) 0.18, Rf1 4 b ) 0.2, Rf1 7 a ) 0.23, and Rf1 7 b ) 0.25. All
tracers exhibited UVmax ) 492 nm in a 0.1 M NaHCO3 solution,
and spectra were identical to that of fluorescein. The ninhydrin
test in all cases was negative except in tracers 1 6 a,b. The
structure of tracers 1 3 a and 1 4 a was confirmed with NMR
spectroscopy. 13a 1H NMR (MeOH-d4): δ ) 1.40 (m, 2H,
CH2-γ), 1.64 (m, 2H, CH2-δ), 1.84 (td, 2H, CH2-â), 3.56 (b, 2H,
CH2ꢀ), 4.34 (b, 1H, CH-R), 4.58 (s, 2H, OCH2CO), 6.51 (dd, 2H,
H-q), 6.64 (d, 2H, H-r), 6.71 (d, 2H, H-p), 7.02 (d, 1H, H-y), 7.08
(d, 1H, H-o), 7.23 (dd, 1H, H-z), 7.38 (d, 1H H-x), 7.70 (d, 1H,
H-m), 8.07 (s, 1H, H-n). 14a 1H NMR (MeOH-d4): δ ) 1.46 (m,
4H, CH2-γ,γ′), 1.62 (m, 4H, CH2-δ,δ′), 1.81 (m, 2H, CH2-â), 1.86
(m, 2H, CH2-â′), 3.59 (m, 4H, CH2-ꢀ,ꢀ′), 4.32 (t, 1H, CH-R), 4.48
(t, 1H, CH-R′) 4.62 (s, 2H, OCH2CO)), 6.53 (dd, 4H, H-q), 6.62
(d, 4H, H-r), 6.78 (m, 4H, H-p), 6.96 (d, 1H, H-y), 7.05 (m, 2H,
H-o), 7.19 (dd, 1H, H-z), 7.33 (d, 1H H-x), 7.70 (d, 2H, H-m), 8.07
Synthesis and Molar Ratio Calculation of 2,4,5T Immunogens
2,4,5T-BSA (19b), 2,4,5T-âAla-BSA (20b), and 2,4,5T-ꢀAca-BSA
(21b). Immunogen syntheses were performed with a modified
carbodiimide method. Briefly, 7.65 × 10-3 mmol of each hapten
was activated with 0.01 mmol of NHS and 0.01 mmol of DCC in
100 µL of DMSO. After DCU precipitation the supernatants were
added to a 1.2 mL solution of 14 mg of BSA in DMSO/ carbonate
buffer (0.05 M; 1:1). After 24 h of reaction, the solutions were
dialyzed against water and stored at -20 °C. The molar ratio
calculation of the immunogens (Figure 1) was performed accord-
ing to Franek.18 The 2,4,5T-BSA (1 9 b) molar ratio ) 20:1, 2,4,-
5T-âAla-BSA (2 0 b) molar ratio ) 13.5:1, and 2,4,5T-ꢀAca-BSA
(2 1 b) molar ratio ) 24/ 1.
Immunization and Antiserum P roduction. Immunization
of six (two for each immunogen) New Zealand rabbits were
carried out by intradermal administration of immunogens with
Freund’s complete adjuvant as elsewhere described.30 Blood
samples were collected 10 days after the boosts. The antiserum
was fractionated with 40% saturated ammonium sulfate solution
and the pellet washed twice with 40% saturated ammonium sulfate
and stored as suspension at 4 °C until use.
FP IA P rocedures. All dilutions of tracers and antibodies were
performed in TDx Abbot buffer. Tracer titer was chosen as the
tracer concentration with a fluorescence intensity of approximately
1000. Final concentrations of all 2,4D tracers used are listed in
Table 1.
Eight serial dilutions for each antibody were prepared, starting
from 1:200. Antibodies dilution curves were obtained by adding
(29) Pourfarzaneh, G. W.; White, J.; Landon, J.; Smith. D. S. Clin. Chem. 1 9 8 0 ,
26 (6), 730-3.
(30) Hatzidakis, G.; Katrakili, K.; Krambovitis, E. J. Reprod. Fertil. 1 9 9 3 , 98,
235-40.
2516 Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

250 µL of tracer (titer dilution), 150 µL of each of the antibody
dilutions, and 300 µL of TDx buffer. The polarization was
measured just after mixing of the reagents. The antibody dilution
with a polarization 30-40% below the maximum value was chosen
as the antibody titer.8
Standard solutions of 2,4D and 2,4,5T were prepared from
methanolic stock solutions (1 mg/ mL) in bottled water with
concentrations 0.1, 1.0, 10, 100, 1000, and 10000 ng/ mL.
Standard curves were obtained using 250 µL of tracer (titer
dilution), 150 µL of antibody (titer dilution), 100 µL of standard
solution of analyte, and 200 µL of TDx buffer. The polarization
measurements were performed after 5 min of incubation at RT
(room temperature).8 The I50 values were calculated using the
Microcal Origin software with sigmoidal fitting.
Figure 2. Schematic representation of the structural homology
appearing between immunogen and tracers 12a,b and the structural
heterology between immunogen and tracers 13a,b. Immunogen and
tracers 12a,b have the antigen (2,4D or 2,4,5T) attached to the amino
group of the ꢀ-primary aliphatic carbon atom of L-lysine. Tracers 13a,b
have the antigen attached to the aminogroup of the R-secondary
asymmetric carbon atom of L-lysine.
Cross- reactivity studies were performed with the most closely
structurally related compounds to evaluate the influence of new
tracers on the specificity of FPIA.
Dissociation constants were calculated with Schatchard plot
analysis according to Dadlinker31 using polarization data.
Apparatus. Fluorescence intensity and polarization measure-
ments were performed with the TDx Abbot fluorometer in
photocheck mode. 1D (1H, 13C) and 2D (gs-COSY, gs-HMQC, gs-
HMBC) NMR spectra were obtained on a Bruker AMX-500
spectrometer using MeOH-d4 as a solvent and at a temperature
of 25 °C. Chemical shifts reported are in ppm from internal TMS.
isolation of high-purity haptens, derivatives, and tracers with
satisfactory reaction efficiencies. The introduction of L-lysine
allowed us to synthesize a variety of monolabeled and bilabeled
tracers. All derivatives exhibited a single spot in the TLC, their
UV spectrum was identical with that of the corresponding analyte,
and the ninhydrin test was positive. The structures of 3 a and 4 a
that were considered as the most important derivatives were
confirmed with 1H NMR spectroscopy, the chemical shifts of
which are given in the Experimental Section. The assignment of
their 1H NMR spectra was performed using 1H-1H gs-COSY, 1H-
13C gs-HMQC, and 1H-13C gs-HMBC 2D NMR techniques
(Figures S1, S2, and S4 in the Supporting Information).
RESULTS AND DISCUSSION
Tracer Selection and Synthesis. For the preparation of
immunogen molecules and subsequent generation of antibodies,
the appropriate hapten molecules are usually attached onto the
surface of a carrier protein through a four-carbon aliphatic side
chain of L-lysine molecules. Antibodies against such immunogens
The very well known final reaction of all amino group
containing derivatives with FITC29 yielded, in all cases, crude
preparations with a main spot in the TLC corresponding to the
expected tracer as confirmed with polarization change measure-
ments using the appropriate antibody. All tracers were purified
with preparative TLC. The structure of tracers 1 3 a and 1 4 a was
confirmed with 1H and 13C 1D and 2D NMR spectroscopy. The
1H NMR chemical shifts of 1 3 a and 1 4 a are given in the
Experimental Section. In the Supporting Information (Figure S3)
often recognize conjugated analytes better than free analytes due
to bridge recognition by the antibody.6,19 This leads to low
sensitivity, as competition between the free and conjugated analyte
(tracer) is favored for the tracer. To address this problem in terms
of FPIA, we prepared a series of 2,4D and 2,4,5T tracers (1 2 a,b,
1 3 a,b, 1 6 a,b) using L-lysine-based derivatives which we consid-
ered as homologous or heterologous to immunogen structure
(Figures 1 and 2). Tracers with two fluorescent molecules
(bilabeled) were also prepared (1 4 a,b, 1 7 a,b) to examine the
potential use of multilabeled tracers and the structure dependency
of quenching in FPIA. Gerdes et al. have used closely related
cross-reactants instead of 2,4D in ELISA conjugates to enhance
competition.32,33 To address this, we prepared tracer 1 5 c (2,4DBr-
EDA-F) on the basis of the cross-reactant 2,4DBr. In addition the
EDA-based tracers 2,4D-EDA-F (1 5 a) and 2,4,5T-EDA-F (1 5 b)
were prepared for comparison purposes, as these tracers were
considered by other researchers as the most appropriate for FPIA
development.8,16,21
1
the H-1H gs-COSY 2D NMR of 1 3 a tracer is given.
Concentration of Tracers: Influence of Quenching. The
concentration of all fluorescent tracers was determined using
ꢀ492 ) 68 mM-1 cm-1 for conjugated FITC.34 All monolabeled
tracers (1 2 a, 1 3 a, 1 5 a, 1 6 a) exhibited approximately 50% lower
fluorescence intensity (FI) when compared with the same con-
centration of free fluorescein, due to quenching caused by hapten-
to-dye interaction13 (Figure S5 in the Supporting Information). The
working concentration (titer) of these tracers in FPIA was set at
2 nM with fluorescence intensity about 1000. In bilabeled tracers
an additional quenching due to dye-to-dye interaction was ob-
served, further reducing the quantum yield. We assumed that this
quenching was structure dependent because tracer 1 4 a exhibited
even lower FI than tracer 1 7 a. This could be explained taking
into account the high flexibility of the two side chains of the
connected lysine moieties and the consequent high possibility of
In Figure 1 the chemical structures of synthesized derivatives
and FITC tracers are shown. The solid-phase synthetic strategy
was favored because it provided advantages in synthesis and
(31) Dandliker, W. B.; Hsu, M.-L.; Levin, J.; Rao, B. R. Methods in Enzymol. 1 9 8 1,
74, 2-28.
(32) Gerdes, M.; Meusel, M.; Spener, F. Anal. Biochem. 1 9 9 7 , 252 (1), 198-
204.
(33) Gerdes, M.; Meusel, M.; Spener, F. J. Immunol. Methods 1 9 9 9 , 223, 217-
26.
(34) Van Dalen, J. P. R.; Haaijman, J. J. J. Immunol. Methods 1 9 7 4 , 5, 103.
Analytical Chemistry, Vol. 74, No. 11, June 1, 2002 2517

Figure 3. Standard curves obtained using the monoclonal antibody E2/G2 and the 2,4D tracers 13a, 15a, 12a, 15c, 16a, 17a, and 14a. In all
cases, mPo was approximately 150. Each standard curve was the mean of four curves obtained from separate experiments. The error bars
presented in the graphs are (1 SD.
interaction between the two fluorescein molecules, as compared
with the interaction between the R- and ꢀ-fluoresceins found in
tracer 1 7 a (Figure 1). The bilabeled tracers 1 4 a and 1 7 a were
used in FPIA with concentrations 8 and 4 nM, respectively. A
tracer with six fluorescein molecules (2,4D-K6(F)6-OH) was also
prepared, which exhibited negligible FI when compared with the
other tracers showing the high degree of dye-to-dye quenching
when the number of dyes/ tracer molecule increases (Figure S5
in the Supporting Information).
2 ,4 D Assay. Standard Curves and Determination of Sensitivi-
ties. For 2,4D FPIA, we used the monoclonal antibody E2/ G2. It
has been also used previously in 2,4D immunoassays.1,24,33,35 In
combination with the 2,4D-EDA-F (1 5 a) tracer, equilibrium and
kinetic FPIAs have been developed.8,16 The I50 was reported as
100 ng/ mL for the equilibrium,8 and the LOD, as 4 ng/ mL for
the kinetic FPIA.16 To compare tracers, the I50 values indicative
of sensitivity were calculated and presented in Table 1. The new
additionally asymmetric and brings a carboxyl group. In our point
of view these structural differences in the site of attachment lead
to more specific recognition of the tracer by the antibody and
limited bridge recognition, and a significant improvement in 2,4D
FPIA sensitivity is reached (I50
) 12.2 ng/ mL). Tracer 1 3 a was
also more sensitive than the bilabeled tracers 1 4 a and 1 7 a, the
sensitivities of which were limited by their high concentration in
reaction mixture due to quenching (Table 1). From comparison
of the bilabeled tracers 1 4 a and 1 7 a, we have found that 1 4 a
exhibits better I50, while this is used in double concentration (8
nM) due to strong dye-to-dye quenching. Tracer 1 4 a is a tether
structure bilabeled tracer, chemically similar to 1 3 a (Figure 1).
This structural similarity is probably the reason for the better
sensitivity obtained when 1 4 a compares to 1 7 a, which is
chemically similar with 1 5 a. Tracers such as 1 4 a and generally
multilabeled tethers can be of great value in assays, which are
not affected by quenching, or in FPIAs with resistant to quenching
dyes. Antigen heterologous tracer 1 5 c, with the 2,4DBr cross
reactant, exhibited surprisingly high I50 (∼154 ng/ mL) suggesting
that important bridge recognition was responsible for the low
sensitivity. Additionally, the weaker recognition of 2,4DBr from
the antibody led to a reduced mPO value. The antibody concentra-
tion was increased 1.5 times in reaction mixture, and an adequate
mPO value (150) was reached. This critical increase in antibody
concentration could be partially responsible for the high I50 value
obtained. This finding also suggests that antigen heterologous
tracers are not probably suitable for FPIA development.
L-lysine-based tracer 2,4D-K(F)-OH (1 3 a) exhibited I50 ) 12.2 ng/
mL and was far more sensitive than all other tracers. The isomeric
tracer F-K(2,4D)-OH (12a) exhibited I50 ) 93.5 ng/ mL, 7.6 times
higher. Tracer 1 3 a was also 5.3 times more sensitive than the
EDA-based one (1 5 a), which exhibited I50 ) 65 ng/ mL. The high
I50 obtained using tracer 1 2 a can easily be explained by taking
into account its total homology with immunogen as Figure 2
shows. 2,4D is attached via the amino group to the ꢀ-primary
aliphatic carbon of L-lysine, in both the immunogen and the tracer
1 2 a. In tracer 1 5 a, 2,4D has been also attached via the amino
group to the primary aliphatic carbon of EDA. This chemical
similarity in the site of attachment between the two tracers is
probably the reason the high I50 also obtained with tracer 1 5 a.
In the case of tracer 1 3 a, 2,4D is attached via the amino group
The standard curves with all tracers and antibody E2/ G2 are
presented in Figure 3. The improvement in sensitivity obtained
with tracer 1 3 a is clearly demonstrated. The dynamic range of
this curve was 1-100 ng/ mL, almost 1 order magnitude better
than the other standard curves. The obtained detection limit was
0.8 ng/ mL, calculated three SD below the mPO value.
to the R-secondary carbon atom of L-lysine. This carbon is
(35) Bauer, C. G.; Eremenko, A. V.; Ehrentreich-Forster, E.; Bier, F. F.; Makower,
A.; Halsall, H. B.; Heineman, W. R.; Scheller, F. W. Anal. Chem. 1 9 9 6 , 68
(15), 2453-8.
Measurement of Dissociation Constant (Kd) Values. To explain
the difference in sensitivity, the Kd values of E2/ G2 antibody with
2518 Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

Figure 4. Scatchard plots from experiments “A” (graphs 1 and 2) and “C” (graphs 3 and 4). The equations given in graphs 1 and 3 correspond
to the scatchard equation BL/FL ) -(1/Kd1)BL + RO/Kd1 for determination of the Kd1 values, when 2,4D is not present in the reaction mixture
(y ) BL/FL, x ) BL, and RO ) total antibody concentration). The equations given in graphs 2 and 4 correspond to the scatchard equation
BA/FA ) -(1/Kd2)BT + RO/Kd2 for determination of the Kd2 value (y ) BA/FA, x ) BT, and RO ) total antibody concentration).
tracers 1 5 a and 1 3 a and with 2,4D were determined in solution
according to Dadlinker et al.31 When 2,4D competes the tracer,
two parallel equilibrium reactions are taking place with different
antibody affinities toward the fluorescent ligand (L) and unlabeled
antigen (A):
higher affinity for the tracer than the free analyte. A sensitive assay
is obtained when the affinity between the tracer and the analyte
is comparable.6 For each tracer two independent experiments with
different antibody quantities each were executed in duplicate. A
and B experiments were performed with tracer 1 5 a, and C and
D experiments with tracer 1 3 a. The results from Scatchard plots
are summarized in Table 2. The Scatchard graphs from experi-
ments A (graphs 1 and 2) and C (graphs 3 and 4) are shown in
Figure 4. In all experiments, the Kd2 values were calculated and
found as expected practically stable, representing the affinity
between the antibody and the free analyte (2,4D). From Kd ratios,
we assumed that the antibody affinity for tracer 1 5 a was much
higher (35 times) than that of free 2,4D. On the contrary, the
affinity for tracer 1 3 a was only 4.3 times higher explaining the
difference in sensitivity (I50) observed between the two tracers.
The ratio in association constants [1/ Kd1(1 5 a):1/ Kd1(1 3 a)] of the
two tracers was 8.1. This difference in affinity can be explained
accepting that there is important bridge recognition in the case
of tracer 1 5 a and negligible bridge recognition for tracer 1 3 a,
explaining its specific nature. The competition of 2,4D was much
R + L a R-L Kd1
R + A a R-A Kd2
R is antibody, L is tracer 1 5 a or 1 3 a, and A is 2,4D. From
Scatchard plots of BL/ FL versus BL when unlabeled antigen (2,4D)
is not present in solution and BA/ FA versus BT in the presence of
2,4D, Kd1 and Kd2 values were respectively calculated. B and F
are the bound and free fractions, respectively, and BT is the total
bound fraction (BT ) BL + BA). Kd2 should be independent of the
tracer used, because it represents the antibody-2,4D dissociation
constant. The difference in tracer structure, the degree of bridge
recognition by the antibody, and the homology or heterology
should lead in tracer-dependent Kd1. Usually the antibody has
Analytical Chemistry, Vol. 74, No. 11, June 1, 2002 2519

Table 2. Dissociation Constants and Relative Affinities
of Tracers 15a and 13a (Kd1 Values) and Herbicide
2,4D (Kd2 values) with the E2/G2 Monoclonal Antibody
Table 4. I50 Values and Their Ratio Obtained from the
3rd and 4th Bleeding of All Immunized Rabbits, Using
the Tracers 13b and 15b
2,4D-EDA-F (1 5 a) 2,4DK(F)-OH (1 3 a)
2,4,5T-K(F)-OH
2,4,5T-EDA-F
(1 5 b)
I50 (ng/ mL)
I50(1 5 b)/
I50(1 3 b)a
ratio
(1 3 b)
constanta
Kd1 (nM)
expt Ab expt Bb
expt Cb
expt Db
antiserum
I
50 (ng/ mL)
0.079
2.84
0.08
2.79
0.64
2.71
0.65
2.83
R1 3rd bleeding
R2 3rd bleeding
R3 3rd bleeding
R4 3rd bleeding
R5 3rd bleeding
R6 3rd bleeding
R1 4th bleeding
R2 4th bleeding
R3 4th bleeding
R4 4th bleeding
R5 4th bleeding
R6 4th bleeding
a-2,4,5T(UK)
226
133
217
85
674
nd
151
212
212
91
385
nd
1028
701
1180
761
2750
4517
793
940
1710
845
1800
3786
1877
4.5
5.3
5.4
8.9
4.1
Kd2 (nM)
1/ Kd1:1/ Kd2
35c
4.3d
1/ Kd1(1 5 a):1/ Kd1(1 3 a)e
8.1
a
1/ Kd is the affinity constant. b Results from two independent
experiments (A, B and C, D) for each tracer. c Ratio of the mean values
of the affinity constants from experiments A and B. d Ratio of the mean
values of the affinity constants from experiments C and D. e Ratio of
the affinity constants between the two tracers.
5.3
4.4
8.1
9.3
4.7
Table 3. % Cross-Reactivities (% CR) of E2/G2
Antibody with 2,4D Structurally Related Compounds,
Using the Tracers 2,4D-K(F)-OH (13a) and 2,4D-EDA-F
(15a)a
389
4.8
a
This is an index of relative sensitivity between the two tracers
indicating the constant improvement obtained with tracer 1 3 b.
% CR with % CR with
2,4D-
K(F)-OH
2,4D-
EDA-F
cross-reactant
Three haptens were synthesized, using solid-phase chemistry,
differing in bridge length: 2,4,5T-BSA (1 9 b), 2,4,5T-âAla-BSA
(2 0 b), and 2,4,5T-ꢀAca-BSA (2 1 b) (Figure 1). With these immu-
nogens six rabbits were immunized, rabbits 1 and 2 with 1 9 b,
rabbits 3 and 4 with 2 0 b, and rabbits 5 and 6 with 2 1 b.
Antiserums after the third bleeding were collected and evaluated
for sensitivity with the corresponding 2,4,5T tracers.
The antiserum from rabbit 3, 3rd bleeding, which exhibited
high titer (1/ 3000), was initially used to evaluate all different 2,4,-
5T tracers. As expected from 2,4D experience, tracers 2,4,5T-K(F)-
OH (1 3 b) and 2,4,5T-EDA-F (1 5 b) gave better results, and
therefore, they were used to evaluate the ammonium sulfate
fractionated antiserums from all immunized rabbits. The resulting
2,4D
2,4,5T
100
3.2
59.8
1.4
1.9
1.7
1.0
0.5
1.2
3.2
100
3.8
60.3
1.4
1.8
2.5
1.2
0.6
6.5
3.2
(2,4-dibromophenoxy)acetic acid (2,4DBr)
2,4-dichlorophenol (2,4D-OH)
(2,4-dimethylphenoxy)acetic acid (2,4DMPA)
(2,5-dichlorophenoxy)acetic acid (2,5D)
(2-chlorophenoxy)acetic acid (2CPA)
(4-chlorophenoxy)acetic acid (4CPA)
(2-chloro-4-fluorophenoxy)acetic acid (2C 4F)
(2,4-dichlorophenoxy)butyric acid (2,4DB)
(2,4-dichlorophenoxy)propionic acid (2,4DP)
2,4-D methyl ester (2,4D-OMe)
<0.1
220.3
<0.1
201.9
a
Cross-reactivities have been calculated using the equation %
CR ) (I50 of 2,4D/ I50 of cross-reactant) × 100.
I
50 values and their ratio from 3rd and 4th bleedings are presented
more effective in the case of tracer 1 3 a giving thus five times
better I50 and consequently better sensitivity in the assay. The Kd
value of 2,4D with E2/ G2 has been measured using an immuno-
sensor24 with absolutely comparable results.
In Table 3 the percent cross-reactivities with the most structur-
ally related compounds are listed. As it can be seen, cross-
reactivities when using tracer 1 3 a were generally similar and in
some cases lower when compared with the ones calculated with
1 5 a tracer. The other characteristics of the assay like recovery,
reproducibility, and accuracy are not practically affected and
remain as previously reported8 using the 2,4D-EDA-F tracer.
in Table 4. Tracer 1 3 b gives in all cases lower I50 values when
compared with 1 5 b. Similar results in ratios were obtained using
another polyclonal antibody against 2,4,5T kindly provided by Dr.
Ramadan Abuknesha from King’s College University of London
with the short name a-2,4,5T(UK). The ratio I50(1 5 b)/ I50(1 3 b)
varies from 4 to approximately 9. The antiserums from rabbits 3
and 4 corresponding to the 2 0 b immunogen with â-alanine as
spacer exhibited the higher ratios. The most sensitive antiserums
were also obtained from rabbit 4, and this was further evaluated.
Antiserum from rabbit 4, 3rd bleeding, exhibited remarkable
sensitivity especially when used with the 1 3 b tracer. Standard
curves from R4 3rd bleeding antiserum and tracers 1 3 b, 1 5 b,
and 1 2 b are provided as Supporting Information (Figure S6). The
other tracers including the bilabeled 14b and 17b exhibited lower
sensitivities due to quenching. The dynamic range of the 1 3 b
tracer curve was from 10 to 1000 ng/ mL with I50 ) 85 ng/ mL
and detection limit at 5 ng/ mL. The standard curves correspond-
ing to the other 2,4,5T tracers appeared with a dynamic range at
least 1 order magnitude higher.
With the use of the absolutely new L-lysine-based tracer 1 3 a,
main characteristics of 2,4D assay have been significantly im-
proved (sensitivity, detection limit) or remained unaffected
(specificity, reproducibility), suggesting the use of this tracer for
enhancing the performance of 2,4D FPIA.
2 ,4 ,5 T Assay. From the above results some questions were
raised having to do with structure 1 3 (Ag-K(F)-OH). (a) Is this a
universal structure that could be adapted and improve other
FPIAs? (b) Can the same results be obtained with polyclonal
antibodies also? (c) Which is the most compatible immunogen
for this structure?
The detailed evaluation of the anti 2,4,5T antibodies is beyond
the aim of this paper. We were focused mainly on the new L-lysine-
based tracer structure performance when used with different
analytes (2,4D and 2,4,5T) and with monoclonal or polyclonal
antibodies derived from different immunogens. We found that,
The herbicide 2,4,5T was used to prepare haptens and
immunogens as already described in the Experimental Section.
2520 Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

in all cases, sensitivity (I50) was improved at least 5 times when
the new tracers 1 3 a,b were used. The ratio in I50 values between
15b and 13b tracers was increased up to 9 times when antibodies
corresponding to immunogen with the three-carbon spacer â-ala-
nine were used.
Franek (Veterinary Research Institute, Brno, Czech Republic) for
the gift of monoclonal antibody E2/ G2, Dr. R. Abuknesha (School
of Health & Life Sciences, King’s College, University of London,
London, U.K.) for the gift of polyclonal antibody a-2,4,5T(UK),
Mikhail Lebedev (Lomonosov Moscow State University) for his
contribution in performing the experimental work, and Dr.
Thanasis Alegakis and Maria Toutoudaki (University of Crete)
for valuable assistance in preparing the manuscript.
CONCLUSION
With the use of L-lysine-based tracers 1 3 a,b we have signifi-
cantly improved the sensitivity of 2,4D and 2,4,5T FPIAs because
the antibody recognition for tracer is specific and the antibody
affinity for tracer is comparable with the affinity for the free
analyte. From our point of view these tracers could be adapted
with success also in other analytes. A number of further improve-
ments including the use of new dyes resistant to quenching, new
instruments, and new technologies, as mentioned in the Introduc-
tion, combined with the tracers presented here could give FPIAs
comparable in sensitivity with that of ELISA. This is of great
importance taking into account the unrivalled simplicity of FPIA.
Our current and future efforts are (a) to apply these materials for
spiked and real sample measurements, like natural water, white
wine, and grape juice, (b) to apply to other FPIA types such as
stopped-flow FPIA, minimizing matrix effect, and further improving
analysis performance, and (c) to apply to a number of other
analytes, including pesticides mycotoxins, drugs, and others.
SUPPORTING INFORMATION AVAILABLE
Figures S1-S6, showing the 1H-1H homonuclear gradient
double quantum filtered COSY-DQF 2D NMR spectrum of
1
derivative 3 a, the H-1H homonuclear gradient double quantum
1
filtered COSY-DQF 2D NMR spectrum of derivative 4 a, the H-
1H homonuclear gradient double quantum filtered COSY-DQF 2D
NMR spectrum of derivative 1 3 a, the carbonyl region of the 1H-
13C heteronuclear gradient, multiple bond correlation HMBC 2D
NMR spectrum of derivative 4 a, the influence of hapten-to-dye
and dye-to-dye quenching in 2,4D tracers, and standard curves
from R4 3rd bleeding antiserum and tracers 1 3 b, 1 5 b, and
1 2 b. This material is available free of charge via the Internet at
http:/ / pubs.acs.org.
Received for review October 1, 2001. Accepted February
19, 2002.
ACKNOWLEDGMENT
The research was partially supported by NATO Collaborative
Linkage Grant (LST.CLG.975077). The authors thank Dr. M.
AC011051X
Analytical Chemistry, Vol. 74, No. 11, June 1, 2002 2521