A new competitive fluorescence assay for the detection of patulin toxin

Article abstract of DOI:10.1021/ac0618526

Patulin is a toxic secondary metabolite of a number of fungal species belonging to the genera Penicillum and Aspergillus. It has been mainly isolated from apples and apple products contaminated with the common storage-rot fungus of apples, Penicillum expa

Full text of DOI:10.1021/ac0618526

Anal. Chem. 2007, 79, 751-757
A New Competitive Fluorescence Assay
for the Detection of Patulin Toxin
Marcella de Champdore´,† Paolo Bazzicalupo,‡ Lorenzo De Napoli,§ Daniela Montesarchio,§
Giovanni Di Fabio,§ Immacolata Cocozza,† Antonietta Parracino,† Mose’ Rossi,† and Sabato D’Auria*,†
Institutes of Protein Biochemistry and of Genetics & Biophysics, CNR, and Department of Organic Chemistry,
University of Naples, Naples, Italy
Patulin is a toxic secondary metabolite of a number of
fungal species belonging to the genera Penicillum and
Aspergillus. It has been mainly isolated from apples and
apple products contaminated with the common storage-
rot fungus of apples, Penicillum expansum, but it has
also been extracted from rotten fruits, moldy feeds, and
stored cheese. Human exposure to patulin can lead to
Figure 1. Chemical structure of patulin.
serious health problems, and according to a long-term
investigation in rats, the World Health Organization has
set a tolerable weekly intake of 7 ppb body weight. The
content of patulin in foods has been restricted to 50 ppb
in many countries. Conventional analytical detection
methods involve chromatographic analyses, such as HPLC,
GC, and, more recently, techniques such as LC/MS and
GC/MS. However, extensive protocols of sample cleanup
are required prior to the analysis, and to accomplish it,
expensive analytical instrumentation is necessary. An
immunochemical analytical method, based on highly
specific antigen-antibody interactions, would be desir-
able, offering several advantages compared to conven-
tional techniques, i.e., low cost per sample, high selec-
tivity, high sensitivity, and high throughput. In this paper,
the synthesis of two new derivatives of patulin is de-
scribed, along with their conjugation to the bovine serum
albumin for the production of polyclonal antibodies.
Finally, a fluorescence competitive immunoassay was
developed for the on-line detection of patulin.
e.g., aflatoxins, ochratoxin A, patulin, and Alternaria toxins.
Mycotoxins may remain in fruits even when the fungal mycelium
has been removed; furthermore, the processing of fruits does not
result in the complete removal of mycotoxins. Human exposure
to mycotoxins occurs by ingestion of contaminated products and
can lead to serious health problems, including immunosuppression
and carcinogenesis.4,5 In animal studies patulin has been shown
to be acutely toxic6,7 and carcinogenic and teratogenic.8 One
important aspect of patulin toxicity in vivo is an injury of the
gastrointestinal tract including ulceration and inflammation of the
stomach and intestine.9 Recently, patulin has been shown to be
genotoxic by causing oxidative damage of DNA,10 and oxidative
DNA base modifications have been considered to play a role in
mutagenesis and cancer initiation.11 Patulin is believed to exert
its cytotoxic effects mainly by forming covalent adducts with
essential cellular thiol groups in proteins and amino acids.12
Indeed, from a chemical point of view patulin is a water-soluble
unsaturated lactone, readily reactive toward thiol groups and also
amino groups. The electrophilic properties of patulin and its
reactivity toward several model nucleophiles, such as N-acetyl-
cysteine and glutathione, have been elucidated.13,14 Glutathione
represents one of the most abundant cellular nucleophiles, and it
Patulin (Figure 1) is a toxic secondary metabolite of a number
of fungal species belonging to the genera Penicillum and Aspergil-
lus. It has been mainly isolated from apples and apple products
contaminated with the common storage-rot fungus of apples,
Penicillum expansum, but it has also been extracted from rotten
fruits, moldy feeds, and stored cheese.1-3 Although several
mycotoxins occur in nature, very few are regularly found in fruits,
(4) CAST 2003; Task Force Report No. 139; Council of Agricultural Science
and Technology: Ames, IA, 2003.
(5) Chu, F. S. Mutat. Res. 1991, 259, 291-306.
(6) Dailey, R. E. J. Toxicol. Environ. Health 1977, 2, 713-725.
(7) Dailey, R. E.; Blasckha, A. E.; Brouwer, E. A. J. Toxicol. Environ. Health
1977, 2, 479-489.
* To whom correspondence should be addressed. Phone: +39-0816132250.
Fax: +39-0816132277. E-mail: s.dauria@ibp.cnr.it.
(8) Sugiyanto, J.; Inouye, M.; Oda, S. I.; Takagishi, Y.; Yamamura, H. Environ.
Med. 1993, 37, 43-46.
(9) Rychlik, M.; Kircher, F.; Schusdziarra, V.; Lippl, F. Food Chem. Toxicol.
2004, 42, 729-735.
(10) Liu, B.-H.; et al. Toxicol. Appl. Pharm. 2003, 191, 255-263.
(11) Olinski, R.; Jaruga, P.; Zastawny, T. H. Acta Biochim. Pol. 1998, 45, 561-
572.
(12) Riley, R. T.; Showker, J. L. Toxicol. Appl. Pharmacol. 1991, 109, 108-126.
(13) Fliege, R.; Metzler, M. Chem. Res. Toxicol. 2000, 13, 373-381.
(14) Fliege, R.; Metzler, M. Chem. Res. Toxicol. 2000, 13, 363-372.
† Institute of Protein Biochemistry, CNR.
‡ Institute of Genetics & Biophysics, CNR.
§ University of Naples.
(1) Buchanan, J. R.; Sommer, N. F.; Fortlage, R. J.; Maxie, E. C.; Mitchell,
F. G.; Hsieh, D. P. H. J. Am. Soc. Hortic. Sci. 1974, 99, 962.
(2) Wilson, D. M. In Advances in chemistry; Rodriks, J. V., Ed.; American
Chemical Society: Washington, DC, 1976; Series 149, p 90.
(3) Bullerman, L. B. J. Food Sci. 1976, 41, 26.
10.1021/ac0618526 CCC: $37.00 © 2007 American Chemical Society
Published on Web 12/10/2006
Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 751

has been shown to spontaneously form covalent adducts with
patulin.13 The same reactivity is responsible for the ability of patulin
to induce in vitro intra- and intermolecular protein cross-links
involving cysteine residues and lysine and histidine side chains15
and for the capability to inhibit the activity of several enzymes.16,17
Patulin is also reported to be endowed with selective DNA-
damaging activity.18
Scheme 1. Synthesis of Derivatives 4 and 5 from
L-Arabinose
The widespread presence of fungi in the environment renders
mycotoxins practically ubiquitous contaminants in food and feeds;
therefore, one of the most effective measures to protect public
health is to establish reasonable regulatory levels of these toxins
on the basis of valid toxicological data. To ensure the safety and
quality of foods, a number of regulatory authorities have decreed
maximum residue levels of several mycotoxins in foods. In
agreement with the results of a long-term investigation in rats,19
the World Health Organization (WHO) has set a tolerable weekly
intake of 7 ppb body weight. The content of patulin in foods has
been restricted to 50 ppb in many countries: the EU has set a
limit of 10 ppb in children’s foods, but the objective is to reach 25
ppb patulin in apple-containing products.20,21 A ready detection is
the most prudent means to prevent the entry of patulin in
commerce. Conventional analytical methods of detection involve
chromatographic analyses, such as HPLC, GC,22,23 and, more
recently, techniques such as LC/MS and GC/MS.24 However,
extensive protocols of sample cleanup are required prior to the
analysis, and expensive analytical instrumentation is necessary
to accomplish it. Furthermore, the problem of precise quantitative
determination of patulin and other mycotoxins by these methods
is particularly important due to the high variability of the real
matrixes and can be solved only by exploiting isotopically labeled
internal standards. All these limitations make it difficult to extend
the methods to detect patulin outside the laboratory. As an
alternative, immunochemical analytical methods, for instance,
enzyme-linked immunosorbent assay (ELISA), offer several ad-
vantages compared to conventional analysis, i.e., low cost per
sample, high selectivity, high sensitivity, and high throughput.
Extensive extraction and sample cleanup are not needed. Such
methods have been shown to have a potential for the routine
analysis of many mycotoxins.25-27 However, no immunoassay has
been developed yet for patulin detection. This is due in part to
chemical reasons, since patulin is, like other nonproteinaceous
toxins, a low molecular weight compound (MW ) 154), and to
develop antibodies, the use of a hapten-protein conjugate is
necessary. In addition, patulin, unlike other toxins, is highly
reactive toward thio-containing compounds and highly unstable
under basic conditions.28 This sensitivity may result in decomposi-
tion of patulin in the course of conjugation to the protein carrier.
Furthermore, the exposed epitopes of the injected hapten-protein
conjugate can be attacked by sulfydryl-containing compounds in
the blood. For this reason few papers have appeared in the
literature on the synthesis of patulin-protein conjugates and on
the production of antibodies against this toxin, and the results
are far from satisfactory in all cases.29-31 A recent review on the
research so far on patulin, from history and regulation to health
effects and existing methods to detect the toxin in foods, has
recently appeared.32
Here the authors describe the synthesis of two new patulin
derivatives, 4 and 5 (Scheme 1), still maintaining the original
skeleton of the natural toxin, but exhibiting a higher chemical
stability, as well as their conjugation to the bovine serum albumin
carrier protein for the production of polyclonal antibodies in
rabbits. They also report on the purification of the produced
antibodies by affinity chromatography on columns obtained by
covalently immobilizing 4 and 5 on Sepharose resin. The titer of
the specific antibodies was determined by means of indirect ELISA
assays. Finally, tetramethylrhodamine isothiocyanate-labeled IgG’s
(IgG ) immunoglobulin G) were prepared and used to develop a
“hit and run” fluorescence immunoassay for patulin, based on the
antibody competition between the immobilized patulin derivative
and free added patulin. This technique was described for the
detection of T-2 toxin33 and more recently for the sensing of 2,4,6-
trinitrotoluene in seawater.34
(15) Fliege, R.; Metzler, M. Chem.-Biol. Interact. 1999, 123 (2), 85-103.
(16) Pfeiffer, E.; Diwald T. T.; Metzler, M. Mol. Nutr. Food Res. 2005, 49 (4)
329-336.
(17) Arafat, W.; Kern, D.; Dirheimer, G. Biochem. Int. 1985, 56 (2-3), 333-
349.
(18) Lee, K. S.; Roschenthaler, R. Appl. Environ. Microbiol. 1986, 152 (5), 1046-
1054.
(19) Becci, P. J.; Hess, F. G.; Johnson, W. D.; Gallo, M. A.; Babish, J. G.; Dailey,
R. E.; Parent, R. A. J. Appl. Toxicol. 1981, 1, 256-261.
(20) WHO, Joint FAO/WHO Expert Committee on Food Additives (JEFCA).
Position paper on patulin, 30th session, The Hague, The Netherlands, 9-13,
1998.
(21) European Commission. EC No. 1425/2003; Brussels, Belgium, 2003.
(22) Lai, C.-L.; Fuh, Y.-M.; Shih, D. Y.-C. J. Food Drug Anal. 2000, 8 (2),
85-96.
(23) Watanabe, M.; Shimitzu, H. J. Food Prot. 2005, 68 (3), 610-612.
(24) Sforza, S.; Dall’Asta, C.; Marchelli, R. Mass Spectrom. Rev. 2005.
(25) Azcona-Olivera, J. I.; Abouzied, M. M.; Plattner, R. D.; Pestka, J. J. J. Agric.
Food Chem. 1992, 40, 531-534.
(28) Harrison, M. A. J. Food Saf. 1987, 9, 147-153.
(29) McElroy, L. J.; Weiss, C. M. J. Can. Microbiol. 1993, 39, 861-863.
(30) Sheu, F.; Lee, O.; Shyu, Y. T. J. Food Drug Anal. 1999, 7 (1), 65-72.
(31) Mhadhbi, H.; Benrejeb, S.; Martel, A. Food Addit. Contam. 2005, 22 (12),
1243-1251.
(26) Huang, X.; Chu, F. S. J. Agric. Food Chem. 1993, 41, 329-333.
(27) Chu, F. S.; Ueno, I. Appl. Environ. Microbiol. 1977, 33, 1125-1128.
(32) Moake, M. M.; Padilla-Zakour, O. I.; Worobo, R. W. Compr. Rev. Food Sci.
Food Saf. 2005, 1, 8-21.
752 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

MATERIALS AND METHODS
µL of SnCl4 was added and the reaction allowed to proceed for 24
h. The reaction mixture was poured into cold H2O (20 mL) and
diluted with CHCl3 (5 mL). The organic layer was separated, and
the aqueous phase was concentrated under reduced pressure,
purified by silica gel chromatography (eluent system: MeOH in
CHCl3 from 20% to 30%), and passed through a DOWEX H+ ion-
All reagents were of the highest commercially available quality
and were used as received. L-Arabinose was purchased from
Sigma. NMR spectra were recorded on Bruker WM-400, Varian
Gemini 300, Varian Gemini 200, and Varian Inova 500 spectrom-
eters. All chemical shifts are expressed in parts per million with
respect to the residual solvent signal. For electrospray ionization
spectroscopy mass spectrometry (ESI-MS) analyses, a Waters
Micromass ZQ instrument, equipped with an electrospray source,
was used in the positive and/or negative mode. 1-[3-(Dimethy-
lamino)propyl]-3-ethylcarbodiimide (EDC), bovine serum albumin
(BSA; fraction V), and ovalbumin (OVA; grade V) were purchased
from Sigma. PURE1A Protein A Antibody Purification Kit was
purchased from Sigma. Goat polyclonal to rabbit IgG-HRP
conjugate (secondary antibody) was from Abcam. Affinity resin
EAH Sepharose 4B was purchased from Amersham Biosciences.
Nitrocellulose transfer membrane Protran from Schleicher &
Schuell and ECL detection reagents from Amersham Biosciences
were used in dot blot and Western blot experiments. Microplates
(96-well), LockWell MaxiSorp from Nunc, 3,3′,5,5′-tetramethyl-
benzidine (TMB) enzyme substrate from Sigma, and a microplate
reader, Multiskan EX from Thermo, were used for ELISA
experiments. UV measurements (detection at λ ) 278 nm) were
carried out on a Varian Cary 50 Bio spectrophotometer. Tetram-
ethylrhodamine isothiocyanate (TMRI) was purchased from
Sigma. Fluorescence experiments were carried out on an ISS K2
fluorometer (ISS, Champaign, IL).
Synthesis and Characterization of Patulin Derivatives.
Synthesis of 4-{[4-(Benzyloxy)-2-oxo-2,6,7,7a-tetrahydro-4H-furo[3,2-
c]pyran-7-yl]oxy}-4-oxobutanoic Acid (3). 4-(Benzyloxy)-7-hydroxy-
7,7a-dihydro-4H-furo[3,2-c]pyran-2(6H)-one (2; 350 mg, 1.37 mmol)
was dissolved in 9 mL of dry THF and treated with 16.7 mg of
DMAP (0.137 mmol) and 1.37 g of succinic anhydride (13.7
mmol), under stirring at room temperature. After 8 h, a second
aliquot of succinic anhydride (685 mg, 6.85 mmol) was added and
the reaction left for an additional 15 h. The reaction mixture was
concentrated under reduced pressure, then redissolved in 25 mL
of ethyl acetate, washed three times with water, and purified by
silica gel chromatography (eluent system: acetone in CHCl3, from
5% to 10%), which resulted in 330 mg (0.91 mmol, 68%) of 3 as a
pure compound. 1H NMR (400 MHz, CDCl3): δH 7.38-7.33 (5H,
complex signals, aromatic protons); 5.98 (1H, d, H-3); 5.70 (1H,
s, H-4); 5.48 (1H, d, H-7a); 5.23 (1H, m, H-7); 4.81 (1H, dd, J )
11.7 Hz, CHHPh); 4.65 (1H, dd, J ) 11.7 Hz, CHHPh); 4.05 (1H,
d, J ) 12.6 Hz, H-6); 3.90 (1H, dd, J ) 12.6 Hz, H-6′); 2.62 (4H, m,
H-10 and H-11). 13C NMR (100 MHz, CDCl3): δC 176.9 (C-9); 171.7
(C-2); 171.0 (C-12); 158.9 (C-3a); 136.1, 128.7, 128.4, and 128.1
(aromatic carbons); 115.0 (C-3); 93.3 (C-4); 77.6 (C-7a); 70.3 (C-
7); 69.7 (CH2Ph); 59.5 (C-6); 28.8 and 28.6 (C-10 and C-11). ESI-
MS m/z calcd for C18H18O8, 362.04; found (positive ions), 385.28
(M + Na+), 401.25 (M + K+).
1
exchange resin to obtain 75 mg (0.28 mmol, 67%) of pure 4. H
NMR (500 MHz, D2O): δH 6.25 (1H, s, H-3); 6.15 (1H, s, H-4);
5.62 (2H, overlapped signals, H-7 and H-7a); 4.36 (1H, d, J ) 13.5
Hz, H-6,); 3.96 (1H, dd, J ) 13.5 Hz and J ) 13.0 Hz, H-6′); 2.67
- 2.57 (4H, m, 4 × H of butanoic acid), 13C NMR (125 MHz,
D2O): δC 178.0 (C-4 of butanoic acid); 173.8 (C-1 of butanoic acid);
165.1 (C-2); 161.7 (C-3a); 114.5 (C-3); 89.0 (C-4); 77.1 (C-7a); 71.6
(C-7); 70.1 (C-6); 29.6 (C-2 and C-3 of butanoic acid). ESI-MS: m/z
calcd for C11H12O8, 272.05; found (positive ions), 255.15 (M - H2O
+ H+), 295.14 (M + Na+), 313.10 (M + K+).
Synthesis of 4-[(4-Hydroxy-2-oxohexahydro-4H-furo[3,2-c]pyran-
7-yl)oxy]-4-oxobutanoic Acid (5, P-sat-HS). To a solution of 3 (150
mg, 0.41 mmol) in 5 mL of MeOH was added a suspension of
150 mg of 10% Pd/C in MeOH (3 mL) under a flux of argon, and
the mixture was hydrogenated at atmospheric pressure for 16 h.
The reaction mixture was filtered through a Celite pad and
concentrated under reduced pressure. The residue was taken up
in 40 mL of H2O and washed twice with DCM. The aqueous layer
1
was concentrated to yield 90 mg (0.32 mmol, 80%) of pure 5. H
NMR (400 MHz, D2O): δH 5.32 (1H, m, H-7); 5.21 (1H, d, J ) 3.2
Hz, H-4); 5.10 (1H, m, H-7a); 4.16 (1H, dd, J ) 12.8 Hz and J )
2.4 Hz, H-6); 3.75 (1H, dd, J ) 13.2 Hz and J ) 1.6 Hz, H-6′); 2.80
- 2.60 (7H, overlapped signals, H-3; H-3a; 2CH2 of butanoic acid).
13C NMR (100 MHz, D2O): δC 183.9 (C-1 of butanoic acid); 183.1
(C-4 of butanoic acid); 177.2 (C-2); 95.9 (C-4); 77.9 (C-7a); 70.1
(C-7); 62.1 (C-3a); 42.7 (C-3); 34.7 (C-2 of butanoic acid); 33.1 (C-3
of butanoic acid). ESI-MS: m/z calcd for C11H14O8, 274.07; found
(positive ions), 297.15 (M + Na+), 313.12 (M + K+).
Synthesis of BSA Conjugates (Antigens A and B). Synthesis
of P-Ins-HS-BSA Conjugate (Antigen A). To a solution of P-Ins-
HS (2 mg, 0.0073 mmol) in 0.5 mL of MES buffer (0.1 M), pH 5,
were added 0.2 mL of a solution of BSA (4 mg‚mL-1) in the same
buffer and 0.1 mL of an EDC solution in H2O (10 mg‚mL-1). The
reaction mixture was incubated for 2 h at room temperature and
then dialyzed against 0.5 L of PBS (0.01 M), NaCl (0.01 M), pH
7.4 (0.5 L, for 3 days with daily buffer changes). The concentration
of the conjugate, spectrophotometrically determined at λ ) 278
nm, was 2.8 mg‚mL-1
.
Synthesis of P-Sat-HS-BSA Conjugate (Antigen B). To a solution
of P-Sat-HS (1.5 mg, 0.0054 mmol) in Tris (pH 8)/dioxane, 1:1
(v/v, 0.4 mL), were added 20 µL (1.0 mg, 0.0054 mmol) of an
EDC solution in H2O (50 mg‚mL-1) and 0.5 mL of a BSA solution
(8 mg‚mL-1) in PBS (0.1 M) at pH 7.4. After 2 h at room
temperature, the reaction mixture was dialyzed against PBS (0.01
M), NaCl (0.01 M), pH 7.4 (0.5 L, for 3 days with daily buffer
changes). The conjugate concentration determined spectropho-
Synthesis of 4-[(4-Hydroxy-2-oxo-2,6,7,7a-tetrahydro-4H-furo[3,2-
c]pyran-7-yl)oxy]-4-oxobutanoic Acid (4, P-ins-HS). To a solution
of 3 (150 mg, 0.41 mmol) in dry DCM (10 mL) was slowly added
SnCl4 (150 µL) at room temperature. After 24 h an additional 75
tometrically at λ ) 278 nm was 4.2 mg‚mL-1
.
Antibody Production and Purification. Antibody Production.
Two rabbits were immunized following a standard protocol by
intradermal inoculation of a mixture of antigens A and B (0.5 mg
each per rabbit). After the immunization period, the rabbits were
(33) Warden, B. A.; Allam, K.; Sentissi, A.; Cecchini, D. J.; Giese, R. W. Anal.
Biochem. 1987, 162, 363-369.
(34) Green, T. M.; Charles, P. T.; Anderson, G. P. Anal. Biochem. 2002, 310,
36-41.
Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 753

sacrificed and the blood was centrifuged to separate blood cells
from serum (SI1 from rabbit 1 and SI2 from rabbit 2).
IgG Purification. A 2.0 mL sample of antiserum (SI1 and SI2)
was applied to a protein A column of the PURE1A Protein A
Antibody Purification Kit, Sigma, and the IgG fraction was purified
according to the manufacturer’s instructions. Elution of proteins
was monitored by absorbance at λ ) 278 nm. The IgG fraction
was eluted with glycine (0.1 M) at pH 2.8 and immediately
buffered in Tris (1.0 M) at pH 8.0. Concentration on an Amicon
XM50 membrane and dialysis against PBS (20 mM), pH 7.0, NaCl
(50 mM) resulted in 4.0 mL of IgG1 and 4.0 mL of IgG2.
Affinity Column Preparation. Two affinity columns were ob-
tained by conjugating derivatives P-Ins-HS and P-Sat-HS, respec-
tively, to EAH Sepharose 4B as follows. A 1.0 mL sample of resin
was washed with H2O at pH 4.5 (20 mL), with NaCl (0.5 M) (20
mL), and again with H2O at pH 4.5 (20 mL) and finally suspended
in 2.0 mL of H2O. The Sepharose resin was added to a solution of
4 or 5 (5 mg in 0.5 mL of H2O at pH 4.5), and the resulting
suspensions were gently shaken. The slurry was cooled to 0 °C,
and EDC was added in two steps to a final concentration of 0.1 M
(52 mg). After 12 h at 4 °C, the reaction mixture was taken to
room temperature, and after an additional 4 h the resin was
extensively washed with H2O at pH 4.5 and then treated with 1.0
mL of AcOH (0.1 M) and 38 mg of EDC for 1 h at room
temperature. The suspension was washed with H2O at pH 4.5 (20
mL), acetate buffer (0.1 M) containing NaCl (0.5 M) (20 mL), pH
4.0, and PBS (0.1 M) containing NaCl (0.3 M), pH 7.4 (20 mL),
and finally packed into a polystyrene column (2 mL, BIORAD).
Antibody Purification by Affinity Chromatography. For the affinity
chromatography purification, a 2.0 mL aliquot of IgG (obtained
from SI1 and SI2 serums) was applied dropwise to the affinity
column prepared as described above. To eliminate unspecific
antibodies, the column, before elution, was washed with three
high-salt buffers: (1) PBS (0.01 M), NaCl (0.1 M), pH 7.0 (20
mL); (2) PBS (0.01 M), NaCl (0.5 M), pH 7.0 (20 mL); (3) PBS
(0.01 M), NaCl (1.0 M), pH 7.0 (20 mL). At that point absorbance
at λ278 nm had fallen to 0.0. For the elution step glycine (0.1 M),
pH 2.7 (2.5 mL), was applied to the column, and the eluate was
collected in 0.5 mL fractions and monitored by absorbance
measurements at λ ) 278 nm.
The fractions containing the antibodies were collected, con-
centrated by means of a Centricon YM-3 membrane to a volume
of 1.0 mL, and dialyzed against PBS (0.1 M), NaCl (0.1 M), pH
7.4. The concentration of the antibodies was spectrophotometri-
cally determined by absorbance measurements at λ ) 278 nm.
Synthesis of GlnBP Conjugates (P-sat-HS-GlnBP, P-ins-HS-
GlnBP). To avoid interference by the carrier protein in the
polyclonal antibody detection process, P-Ins-HS and P-Sat-HS were
conjugated to the glutamine-binding protein from Escherichia coli
(GlnBP), a bacterial protein different from the protein used to
carry out the immunization. The following procedure was used:
2.5 mg (0.0092 mmol) of 4 or 5 was dissolved in 0.25 mL of MES
buffer (0.1 M), pH 5. The solution was incubated at room
temperature with 0.25 mL of a GlnBP solution (5 mg‚mL-1) in
the same buffer and 0.1 mL of an aqueous solution of EDC (10
mg‚mL-1). After 5 h, the reaction mixture was dialyzed against
PBS (0.01 M) containing NaCl (0.1 M) at pH 7.4 (0.5 L, 3 days
with daily buffer changes).
Figure 2. Schematic representation of the BSA conjugates used
for the production of the rabbit anti-patulin antibodies.
Dot Blot Experiments. P-sat-HS-GlnBP, P-ins-HS-GlnBP,
and GlnBP were spotted on four nitrocellulose membranes (10
and 1 µg of each antigen on each filter) and the spots allowed to
dry. The filters were washed twice with PBS (0.1 M), NaCl (0.4
M) at pH 7.4 (PBS) and then incubated with the blocking buffer
(PBS containing 5% skim milk, 0.2% Tween 20, and 0.05% Tryton,
pH 7.4) for 1 h at room temperature. After two washings with
PBS (0.1 M), NaCl (0.4 M), Tween (0.2%), and Tryton (0.05%) at
pH 7.4 (PBS-TT) and one with PBS (10 min per washing), the
filters were incubated with SI1, SI2, SPI1, and SPI2 (each diluted
1:250 in blocking buffer), respectively, for 1 h at room temperature.
The filters were washed (twice with PBS-TT and one time with
PBS, 10 min per washing) before being incubated with the
secondary antibody (goat anti-rabbit HRP conjugate, 1:3000 in
blocking buffer) for 1 h at room temperature. The filters were
washed three times as described above and developed with the
detection reagent ECL.
Western Blot Experiments. Proteins (BSA, P-Ins-HS-GlnBP
or P-Sat-HS-GlnBP, and GlnBP, 10 µg each) were loaded,
separated by sodium dodecyl sulfate-polyacrylamide gel electro-
phoresis (12% SDS-PAGE), and then transferred overnight at 4
°C onto a nitrocellulose membrane. Membranes were blocked for
1 h at room temperature by rocking in 50 mL of the blocking
buffer (PBS containing 5% skim milk, 0.2% Tween 20, and 0.05%
Tryton). After two washings with PBS-TT and one with PBS (10
min per washing), the filters were incubated with a-PiHS and
a-PsHS, respectively (1:500 in the blocking buffer), for 1 h at room
temperature. After two washings with PBS-TT and one with PBS
(10 min per washing), the filters were incubated with secondary
antibody (goat anti-rabbit HRP conjugate, 1:3000 in the blocking
buffer) for 1 h at room temperature. The filters were washed three
times as described above and then developed with the detection
reagent ECL.
Antibody Titration. The antibody titer was determined by an
indirect ELISA assay by the following general procedure. The
antigens (P-Ins-HS-GlnBP or P-Sat-HS-GlnBP), in PBS (0.1 M),
pH 7.4, were used to coat 96-well microplates, varying the
concentration by a factor 3 from 11 to 1.7 × 10-3 µg‚mL-1 (one
column for every antigen concentration, 100 µL per well),
overnight at 4 °C. Control wells were incubated for the same
754 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

diluted 1:4000 in PBS-T containing BSA (1%), were added to the
wells (100 µL) and incubated for 1 h at room temperature. After
three washings with PBS-T, the enzyme substrate TMB was
added (100 µL per well), and the color reaction was quenched
after 5 min by addition of 1 M H2SO4 (100 µL per well). The
absorbance was measured at λ ) 450 nm. The antibody titer was
graphically determined by plotting the reciprocal of the antibody
dilution vs absorbance for each dilution of antibodies. The titer
was taken as the maximum antibody dilution able to give a reading
of 0.1 absorbance unit. The following values were found: 1/25000
for aPsHS1 and aPsHS2 and 1/25000 for aPiHS1 and aPiHS2.
Synthesis of the IgG-TMRI Conjugate. IgG fractions from
each rabbit serum were labeled with TMRI as follows. A 0.7 mL
sample of IgG in buffer, NaHCO3 (10 mM), at pH 9.0 (2.5
mg‚mL-1) was incubated with a solution of TMRI (0.15 mg‚mL-1
)
in DMF (50 µL)/NaHCO3 (10 mM), pH 9.0 (450 µL), for 2 h at
room temperature. The labeled antibodies were purified on a
Sephadex G-25 column. The fractions containing labeled IgG were
pooled and concentrated to a volume of 2.0 mL on a Centricon
microconcentrator. The degree of labeling (DOL), as calculated
from the absorbance values at λ ) 278 and 550 nm, by applying
a correction factor for label absorption at λ ) 278 nm, was found
to be 3.5.
Binding of IgG-TMRI to Patulin Derivatives on Affinity Columns
and Detection of Free Patulin. The previously prepared affinity
column was washed and equilibrated with the binding buffer PBS
(0.1 M), NaCl (0.1 M), pH 7.4. A 1.5 mL sample of diluted IgG-
TMRI (dilution 1:3 in binding buffer) was applied dropwise onto
the resin. The column was washed with the binding buffer (20
mL), then with PBS (0.1 M), NaCl (0.5 M), pH 7.4 (20 mL), and
finally with PBS (0.1 M), NaCl (1 M), pH 7.4, until the fluorescence
signal was returned to the preloading level (λ ) 550 nm excitation,
λ ) 575 nm emission). Samples of patulin (0, 1, 10, and 100 ppb)
in 2.0 mL of binding buffer were applied to the column at room
temperature and incubated for 10 min. At the end of the incubation
period, the patulin-IgG-TMRI complex was eluted from the
column and collected in 0.5 mL fractions. The fluorescence
emission signal was measured at λ ) 575 nm. The column was
washed with 5.0 mL of the binding buffer and could be reused
for the analysis of the successive patulin sample.
Figure 3. SDS-PAGE and Western blot analyses of the specific
antibody anti- derivative 5. SDS-PAGE stained with Coomassie
Brillant Blue R-250 (A) and nitrocellulose filter (B) after incubation
with RPsHS1 antibodies. Key: lane 1, molecular weight standards;
lane 2, bovine serum albumin; lane 3, GlnBP; lane 4, P-Sat-HS-
GlnBP.
DISCUSSION AND RESULTS
Synthesis of Patulin Derivatives 4 and 5. The presence of
a carrier protein conjugated to the toxin is required for the
production of antibodies against patulin, which is, like other
nonproteinaceous toxins, too small to elicit any immunological
response. However, the conjugation reaction is hampered by the
reactivity of patulin functional groups in the reaction conditions
required to functionalize the OH group at the C-4 position (Figure
1). Indeed, from a chemical point of view patulin is a highly
conjugated bicyclic lactone and, in general, very reactive at the
C7 position under basic conditions and toward nucleophiles,
especially thiol and amino groups.13 Furthermore, even if the
desired antigen could be obtained, the patulin epitopes would be
sensitive, in vivo, to the SH-containing molecules in the blood,
thus rendering the intact molecule unavailable and the recognition
process disfavored, with no or poor production of antibodies. With
the aim of reducing the reactivity of patulin toward nucleophiles,
two new derivatives, 4 (P-Sat-HS) and 5 (P-Ins-HS) (Scheme 1),
Figure 4. Chemical structure of tetramethylrhodamine isothiocy-
anate.
period with BSA in the same buffer. The wells were rinsed three
times with PBS (0.1 M) containing 0.05% Tween (PBS-T), pH
7.4, and blocked by incubation for 2 h at room temperature with
PBS-T containing BSA (1%) (100 µL each well). After three
washings with PBS-T, serially diluted antibody, aPiHS1, aPsHS1,
aPiHS2, or aPsHS2, was added to the wells, incubated at room
temperature for 1 h, and then rinsed three times with PBS-T.
Horseradish peroxidase-conjugated anti-rabbit IgG antibodies,
Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 755

Figure 5. Fluorescence immunoassay for patulin detection. Fluorescence emission was recorded at λ ) 575 nm and is representative of the
IgG-TMRI eluted from the column after incubation of patulin. The fluorescence experiments were performed at room temperature; excitation
was at 550 nm.
lacking the highly reactive C7-C7a double bond, were synthe-
sized starting from compound 3 (Scheme 1). This in turn was
obtained from derivative 2, prepared from commercial L-arabinose
antiserums (SI1 and SI2, 1:250 dilution) and preimmune serums
from both rabbits (SPI1 and SPI2, respectively, from rabbit 1 and
rabbit 2, dilution 1:250). After incubation of the HRP conjugate
secondary antibody and development with ECL-specific reagent,
preimmune serums showed no response, while SI1 and SI2 gave
signals with both antigens P-sat-HS-GlnBP and P-ins-HS-GlnBP
and with BSA but not with GlnBP (not shown).
Immunoglobulin Purification from Serums. The IgG frac-
tion of each serum (IgG1 from rabbit 1 and IgG2 from rabbit 2)
was isolated by the protein A column kit by standard procedures.
After elution, the IgG fraction was concentrated and dialyzed
against PBS (20 mM), NaCl (50 mM), pH 7.0.
Preparation of Affinity Columns and Purification of Spe-
cific Antibodies. Antibodies specific for either compound 4 or
compound 5 were purified from the IgG fraction by affinity
chromatography on one of two resins conjugated with compound
4 or 5. After loading, the columns were washed with high-salt
buffers to eliminate unspecific antibodies and eluted using a buffer
at pH 2.7. The fractions containing the antibodies, as judged by
monitoring the absorbance at λ ) 278 nm, were pooled and tested
against the patulin derivatives by Western blot. Only antibodies
eluted by changing the pH conditions were effective in the
recognition of compounds 4 and 5, as shown in Figure 3 (aPsHS
and aPiHS, respectively). These antibodies were used in the
experiments of antibody titration (aPsHS1 and aPiHS1 from rabbit
1 and aPsHS2 and aPiHS2 from rabbit 2). To the best of our
knowledge, this is the first time that polyclonal antibodies
produced against patulin have been purified by affinity chroma-
tography, due to the chemical stability of derivatives 4 and 5,
which allowed them to be conjugated to an NH-functionalized
by a previously described synthetic scheme.35 Removal of the
benzyl protecting group in 3 by treatment with SnCl4 yielded
compound 4; conversely, catalytic hydrogenation of 3 nicely
afforded compound 5. In both cases the synthesis gave a mixture
of diastereoisomers of 4 and 5, which were not separated and
used as such in the successive protein conjugation reactions.
Conjugation of Patulin Derivatives to the Bovine Serum
Albumin. BSA was chosen as the carrier protein, and haptens 4
and 5 were conjugated to it by the water-soluble carbodiimide
method, leading to antigens A and B (Figure 2). The efficiency
of the conjugation reaction could not be determined because of
the lack of a UV chromophore in the case of derivative 5 and
because the maximum absorption at λ ) 278 nm of 4 and BSA
overlapped.
Production of Antibodies against the Patulin Derivatives.
Both antigens A and B were used to immunize rabbits for the
production of polyclonal antibodies against patulin, following a
standard protocol. At the end of the immunization period, the
presence of antibodies against patulin derivatives in serums (SI1
serum from rabbit 1 and SI2 serum from rabbit 2) was verified
by dot blot assay.
To exclude false reactions due to anti-BSA antibodies present
in the rabbit serums, a protein different from BSA was conjugated
to the patulin derivatives P-Sat-HS and P-Ins-HS. For this purpose,
the glutamine-binding protein (GlnBP) from E. coli was chosen.
The GlnBP conjugates were prepared by the same procedures
already used for BSA conjugates.
P-sat-HS-GlnBP, P-ins-HS-GlnBP, BSA as the positive control,
and GlnBP as the negative one were spotted on four nitrocellulose
membranes, and each membrane was incubated with immune
(35) Bennett, M.; Gill, G. B.; Pattenden, G.; Shuker, A. J.; Stapleton, A. J. Chem.
Soc., Perkin Trans. 1 1991, 929-937.
756 Analytical Chemistry, Vol. 79, No. 2, January 15, 2007

resin. The affinity columns were also used to develop the
fluorescence competitive immunoassay for the detection of patulin.
Specific Antibody Titration. The titer of purified antibodies
was determined by indirect ELISA, by coating on the microplate
wells several different concentrations of antigens P-sat-HS-GlnBP
and P-ins-HS-GlnBP and by testing serially diluted aPsHS1 and
aPsHS2 against P-sat-HS-GlnBP and P-ins-HS-GlnBP, respec-
tively. Each experiment was performed in triplicate, and the results
showed that the titer of antibodies, expressed as the reciprocal
dilution giving 0.1 optical density (OD) unit at λ ) 450 nm, was
25000 for all the tested antibodies (data not shown).
Patulin Detection by Affinity Chromatography with IgG-
TMRI. A hit and run fluorescence immunoassay33 was performed
by labeling the IgG fraction with the fluorophore TMRI (Figure
4). The technique is based on the binding of fluorescence-labeled
IgG to the derivative P-sat-HS or P-ins-HS, immobilized on a
Sepharose 4B matrix, as previously described, followed by
displacement of the TMRI-labeled antibodies by patulin. Displaced
fluorescence antibodies are measured by fluorescence emission
in the eluate and represent a measure of the amount of free patulin
present in the sample. The labeling of the IgG fraction was
achieved by incubating it with TMRI (previously dissolved in a
mixture of DMF and binding buffer, pH 9.0). After purification of
the labeled IgG by gel filtration, an aliquot of IgG-TMRI was
applied to the column previously loaded with patulin derivative 4
or 5. The column was washed with buffer, and the fluorescence
emission was monitored at λ ) 575 nm (excitation at λ ) 550
nm). When the signal returned to the preloading IgG-TMRI level,
samples of patulin (0, 10, and 100 µg‚L-1) in 2.0 mL of binding
buffer, PBS (0.1 M), NaCl (0.1 M), pH 7.4, were applied to the
column and incubated for 10 min at room temperature: 0.5 mL
fractions were collected, and the fluorescence emission was
measured, providing a measure of the IgG-TMRI complex
released and, in turn, the amount of free patulin present in the
sample. Each experiment was performed in triplicate on each
affinity column. Representative peaks of fluorescence are shown
in Figure 5 for the experiments performed with the column on
which P-Ins-HS was immobilized. Similar results were obtained
for the derivative P-Sat-HS. A concentration of patulin of 10 µg‚L-1
is easily detectable by this technique. To test the specificity of
the method, the antibiotic neomycin was also incubated with the
matrix-IgG-TMRI complex under the same conditions used for
patulin. In this case, as expected, no fluorescence emission at λ
) 575 nm was detected.
CONCLUSIONS
The polyclonal antibodies produced against the mycotoxin
patulin, by means of two synthetic derivatives of the molecule
conjugated to bovine serum albumin as the carrier protein, after
affinity purification, showed a high titer for the recognition of the
haptene. The immunoglobulins were successfully used to develop
a competitive immunofluorescent assay for the detection of small
quantities of patulin. The method was based on the binding of
fluorescently labeled antibodies to the synthetic patulin derivative,
covalently immobilized on a Sepharose matrix. The displacement
of the antibodies by addition of free patulin was detected as a
fluorescence signal at the emission wavelength of the fluorophore.
The assay allowed the detection of 10 µg/L patulin and thus
represents a promising basis for the development of an alternative
methodology compared to the analytical chromatographic tech-
niques in use.
ACKNOWLEDGMENT
This work was supported by the ASI project MoMa, No. 1/014/
06/0, by a grant from the Ministero degli Affari Esteri, Direzione
Generale per la Promozione e la Cooperazione Culturale (S.D.,
M.d.C.), and by the CNR Commessa Diagnostica Avanzata ed
Alimentazione (S.D.).
Received for review October 3, 2006. Accepted November
9, 2006.
AC0618526
Analytical Chemistry, Vol. 79, No. 2, January 15, 2007 757