Bioseparation 10: 389394, 2002.

2002 Kluwer Academic Publishers. Printed in the Netherlands.

389
A molecular imprinted polymer with recognition properties towards the
carcinogenic mycotoxin ochratoxin A
Claudio Baggiani

, Gianfranco Giraudi & Adriano Vanni

Dipartimento di Chimica Analitica, Universit ` a degli Studi di Torino, Via Giuria 5, 10125 Torino, Italy (
Corresponding author; E-mail: claudio.baggiani@unito.it)
Key words: molecular imprinting, ochratoxin A, mycotoxin, liquid chromatography, molecular recognition
Abstract
A molecularly imprinted polymer which recognises the mycotoxin ochratoxin A was prepared using the mimic
N-(4-chloro-1-hydroxy-2-naphthoylamido)-(L) -phenylalanine as a template. The polymer was obtained by dis-
solving the template, methacrylic acid and ethylendimethacrylate in chloroform and polymerising the mixture by
thermal treatment at 60

C. The monolith obtained was crushed, sieved to 3090 m and extensively washed
till the template could no longer be found in the washing solution. The binding properties towards the template,
ochratoxin A and several related molecules were measured by eluting with acetonitrile and chloroform a HPLC
column packed with the imprinted polymer. The experimental results show that the polymer recognises not only
the template well, but also the ochratoxin A. The specic molecular recognition effect is due to hydrogen bond
interactions but in order to assure the full recognition effect adjunctive steric factors are necessary. The magnitude
of these interactions can be controlled by the use of limited amounts of acetic acid in the mobile phase.
From the measurement of the relative selectivity it was found that only the simultaneous presence of the
carboxyl, the phenolic hydroxyl and certain peculiar substructures such as the chlorine atom assures the whole
recognition of the template.
Abbreviations: amido-CHNA 4-chloro-1-hydroxy-2-naphthoylamide; CHNA 4-chloro-1-hydroxy-2-
naphthoic acid; D-Phe-CHNA N-(4-chloro-1-hydroxy-2-naphthoylamido)-(D) -phenylalanine; HNA 1-
hydroxy-2-naphthoic acid; k retention factor; L-Ala-CHNA N-(4-chloro-1-hydroxy-2-naphthoylamido)-(L)
-alanine; L-Gly-CHNA N-(4-chloro-1-hydroxy-2-naphthoylamido)-glycine; L-Phe-CHNA N-(4-chloro-
1-hydroxy-2-naphthoylamido)-(L) -phenylalanine; L-Phe-CSA N-(4-chlorosalycylamido)-(L)-phenylalanine;
L-Phe-HNA N-(1-hydroxy-2-naphthoylamido)-(L)-phenylalanine; L-Phe-NA N-(2-naphthoylamido)-(L)-
phenylalanine; L-PheOMe-CHNA N-(4-chloro-1-hydroxy-2-naphthoylamido)-(L) -phenylalanine methyl ester;
L-Trp-CHNA N-(4-chloro-1-hydroxy-2-naphthoylamido)-(L) -tryptophane; L-Tyr-CHNA N-(4-chloro-1-
hydroxy-2-naphthoylamido)-(L) -tyrosine; MIP molecular imprinted polymer; NIP nonimprinted polymer;
phenethyl-CHNA N-(4-chloro-1-hydroxy-2-naphthoylamido)-phenethylamine; t retention time of the eluted
substance; t
0
retention time corresponding to the column void volume; tyramine-CHNA N-(4-chloro-1-
hydroxy-2-naphthoylamido)-tyramine.
Introduction
The mycotoxin ochratoxin A, N-(R)-phenylalanine-
5-chloro-3,4-dihydro-8- hydroxy-3-methyl-1-oxo-1H-
2-benzopyran-7-amide, is a naturally occurring my-
cotoxin, a metabolite produced by several widely dif-
fused fungi such as Aspergillus ochraceus, sulphureus
and melleus in temperate climates, and Penicillium vi-
ridicatum in tropical and subtropical areas. It is well
known to be a natural contaminant on cereals, rice,
peanuts, coffee beans, cottonseed and decaying ve-
getation. Residues have been detected in samples of
our, bread, sausage, ham and bacon samples, as well
as meat from animals slaughtered immediately after
consuming contaminated feed. This substance is reas-
onably anticipated to be a human carcinogen based
390
on sufcient evidence of liver and kidney carcinomas
induced in experimental animals fed with it. It is con-
sidered to be also the primary cause of the so-called
Balkan endemic nephropathy, a degenerative affec-
tion of kidney widely diffused in several east-european
countries (Poland et al., 1982, 1992; US Dept. Health,
2000).
For these reasons, food contamination by ochratoxin
A is a growing concern for public health, and the
trace detection in dairy products and food samples is
an analytical problem of primary importance for the
food industry. The need to perform a large number
of analysis on complex samples leads the analyst to
use selective extraction and preconcentration methods,
i.e. immunoafnity columns, in spite of their high
costs and short life (Nakajima et al., 1997; Scott &
Trucksess, 1997; Wilkes & Sutherland, 1998).
Polymeric stationary phases obtained by the tech-
nique of molecular imprinting could be a valid al-
ternative to the immunoafnity phases in terms of
limited costs, column stability and reproducibility
(Olsen et al., 1998; Sellergren, 1999; Stevenson,
1999). Nevertheless, the preparation of an ochratoxin
Aimprinted polymer poses several practical problems:
the template molecule is very expensive, there are
severe safety problems and the manipulation of sev-
eral hundreds of milligrams is not easily affordable
for an ordinary research laboratory. Last but not least,
ochratoxin Ais degraded by thermal or UVirradiation,
conditions normally used to polymerise the template
mixture.
The use of a template that mimics the structure
of a related molecule and acts successfully as a pu-
tative imprinting molecule has been reported in the
literature (Andersson et al., 1997; Matsui et al., 2000a,
2000b). The use of the so-called dummy template
polymerisation technique appears to be the solution
to the problems described here. Moreover, a polymer
imprinted with a mimic molecule is suitable for ap-
plications in the eld of solid phase extraction at trace
level, solving the problem of template bleeding that
could affect these kind of materials (Andersson et al.,
1997; Rashid et al., 1997).
In this work we describe the preparation and the
characterisation of a molecular imprinted polymer
binding ochratoxin A obtained by using a mimic mo-
lecule able to raise specic binding sites during the
polymerisation process, thus similar to ochratoxin A,
but easy and economical to prepare, stable to heat and
UV light, and less dangerous to health.
Experimental
Materials
Ochratoxin A, technical grade, was obtained from
Fluka (Milan, Italy) and used without further puric-
ation. All others reagents were from SigmaAldrich
(Milan, Italy). Ethanol-free chloroform used as poro-
gen solvent was obtained from commercial HPLC-
grade chloroformby distillation. Methacrylic acid and
ethylene glycol dimethacrylate were distilled at re-
duced pressure immediately before use. An ochratoxin
A stock solution was prepared dissolving 2.5 mg of
mycotoxin in 1.0 ml of anhydrous acetonitrile and
storing in the dark at 20

C.
The molecular structures of ochratoxin A, dummy
template and related analogs were obtained with Hy-
perchem 5.01 (Hypercube, Gainesville, FL, USA).
The structures were minimised using Hyperchems
mm+ molecular mechanic method, then were rened
using PM3 as semiempirical method. Approximate
solvent accessible surface, electrostatic potential sur-
face and lypophilic/hydrophilic surface were visual-
ised with Chime 2.0 (MDL Information Systems, San
Leandro, USA).
HPLC apparatus (pump L-6200, UVVIS detector
L-4200 and integrator D-2500) was from Hitachi
Merck (Darmstadt, Germany). Reverse-phase HPLC
column (C18 Macrosphere, 250 4.6 mm) were from
Alltech (Milano, Italy).
Synthesis of 4-chloro-1-hydroxy-2-naphtoic acid
The chlorination of the commercially available 1-
hydroxy-2-naphthoic acid was carried out according
to the procedure indicated in literature (Airan, 1942).
In a 250-ml round-bottom ask provided with reux
condenser and dry trap 5.0 g of HNA were suspended
in 50 ml of anhydrous diethylether and gently heated
till completely dissolved. A catalytic amount of bis-
muth trichloride was added and a solution of 2.13 ml
of sulfuryl chloride in 10 ml of anhydrous diethylether
was added drop by drop under continuous stirring.
Then the mixture was poured into 50 ml of deionised
water and the organic layer slowly evaporated under a
stream of nitrogen. The white-brown precipitate ob-
tained was recrystallised twice in absolute ethanol,
giving the pure product as a white powder (3.2 g, 54%
yield), deemed pure by reverse-phase liquid chroma-
tography (mobile phase: methanolwateracetic acid,
85:14:1, v/v). FT-IR (KBr, cm
1
): 3072, 1590, 1476
391
(naphthalene), 1676 (carboxyl), 1385 (phenol). Mass
spectrum API-ES (m/z, relative intensity): 221 (M
+
,
100), 177 (M COOH, 62).
Synthesis of
N-(4-chloro-1-hydroxy-2-naphthoylamido)-(L)
-phenylalanine and related derivatives
The synthesis of the dummy template and sev-
eral related derivatives was developed utilising the
reaction between a primary amine and the N-
hydroxysuccinimidyl derivative of a carboxylic acid.
For example, here the synthesis of the dummy tem-
plate is reported.
In a 50-ml round-bottom ask, 1.0 g of CHNA,
1.02 g of N, N-dicyclohexylcarbodiimide and 0.569
g of N-hydroxysuccinimide were dissolved into 30
ml of anhydrous tetrahydrofuran. The mixture was
stirred overnight at room temperature, the N, N-
dicyclohexylurea formed was separated by ltration in
a G4 Buchner, and then 0.877 g of L-phenylalanine
dissolved in 30 ml of sodium hydrogencarbonate 0.15
m were added. The mixture was stirred overnight,
then evaporated in a rotavapor. The residue, a yellow-
white solid, was dissolved into 50 ml of dichloro-
methane, then washed three times with 20 ml of
10 mm aqueous hydrogen chloride and three times
with 50 ml of water. The organic layer was dried
over anhydrous sodium sulphate and evaporated un-
der a stream of air. The raw powder obtained was
recrystallised twice in absolute ethanol, giving the
pure product as a white powder (1.33 g, 80% yield),
deemed pure by reverse-phase liquid chromatography
(mobile phase: methanolwateracetic acid, 85:14:1,
v/v). FT-IR (KBr, cm
1
): 3080, 1600, 1490 (naph-
thalene) 3005 (aromatic), 1720 (carboxyl), 1680, 1560
(amide), 1380 (phenol). Mass spectrum API-ES (m/z,
relative intensity): 368 (M
+
, 11), 324 (M COOH, 8),
221 (M Phe, rearranged, 26), 177 (CHNA COOH,
100), 141 (HNA COOH, 4).
Polymer preparation
In a 10-ml test tube a solution was prepared by dis-
solving 0.100 g (0.270 mmol) of L-Phe-CHNA in 4.8
ml of chloroform. Then, 0.162 ml of methacrylic acid
(1.62 mmol), 2.85 ml of ethylene glycol dimethac-
rylate (14.6 mmol) and 0.040 g of 2,2-azobis-(2-
methylpropionitrile) were added. The mixture was
purged with nitrogen and sonicated in a water-bath
for 5 min. The tube was sealed and the mixture was
left to polymerise overnight at 60

C. The polymer ob-

tained, a white, solid mass, was broken with a steel
spatula, mechanically ground in a mortar and wet-
sieved to 3090m particle size. The particulate was
extensively washed with methanolacetic acid (9:1
v/v). No efforts were made to measure the amount of
template molecule recovered. A blank polymer (NIP)
was prepared and treated in the same manner, omitting
L-Phe-CHNA.
Liquid chromatography
An adequate amount of polymer was suspended in
acetonitrile, sonicated in a water-bath for 10 min and
the slurry packed in a 100-mm stainless-steel HPLC
column (I.D. 3.9 mm, geometrical volume 1.19 cm
3
).
The packing of the stationary phase was performed
by gradually adding the slurry of the polymer to the
column and eluting it with acetonitrile at constant
pressure of 10 MPa. The packed column was washed
at 0.5 ml/min with methanolacetic acid (9:1 v/v) until
a stable baseline was reached (310 nm). After equilib-
ration, the pressure in the column was 13 MPa using
acetonitrile as a mobile phase and at a ow rate of 0.5
ml/min.
The column was equilibrated at a ow rate of 0.5
ml/min with 40 ml of proper mobile phase; then, 20
l of stock solution of ochratoxin A (or a related
substance) diluted 1:50 (v/v) with acetonitrile were
injected and eluted at 0.5 ml/min, and the absorb-
ance recorded at 310 nm. Each elution was repeated
three times to assure the chromatogram reproducib-
ility. Column void volumes were measured for each
mobile phase formulation by eluting 20 l of acetone
0.1%(v/v) in acetonitrile, and the absorbance recorded
at 240 nm.
The retention factor was calculated as (t t
0
)/t
0
.
The selectivity factor (index of polymer selectivity
towards analogues of the template molecule) was
calculated as k
analogue
/k
template
.
Results and discussion
Choice of the template
A good mimic of ochratoxin A suitable for a success-
ful imprint should preserve the general structure of the
molecule, including the chirality of the aminoacidic
sub-structure and the planarity of the benzopiranic
sub-structure. At the same time, it is necessary to elim-
inate the -unsaturated lactone moiety, to which the
392
Figure 1. Molecular structures of ochratoxin A (a) and of the mimic
L-Phe-CHNA (b).
carcinogenicity of many known mycotoxins is related.
Moreover, to assure an efcient imprinting effect the
several distinct points of potential interaction with
monomers should be maintained: the -carboxyl of
L-phenylalanine, the amido bridge, and the phenolic
hydroxyl.
Among many possible template structures, L-Phe-
CHNA was chosen because of its simplicity of syn-
thesis. Actually, this molecule could be obtained from
cheap and commercially available substances through
two simple and high yield reactions. Other possible
templates, such as those derived by partially satur-
ated naphthalene rings, revealed themselves to be too
difcult or expensive to be synthesised. A prelimin-
ary study performed on the molecular structures of
ochratoxin A and of L-Phe-CHNA (Figure 1) also
showed a quite complete overlapping of the two mo-
lecules, with a high degree of similarity not only as
structures, but also as solvent accessible surfaces, elec-
trostatic potential surfaces and lypophilic / hydrophilic
surfaces.
Effect of mobile phase
The retention properties of the imprinted and blank
columns were evaluated by eluting the template and
the ochratoxin Awith acetonitrile and chloroformcon-
taining different amounts of acetic acid, ranging from
0.05 to 1% in volume (Figures 2 and 3). The im-
printed column shows capacity factors signicantly
Figure 2. Retention of ochratoxin A (squares) and L-Phe-CHNA
(circles) on the imprinted (closed points) and blank (open points)
columns eluted with chloroform containing variable amounts of
acetic acid.
Figure 3. Retention of ochratoxin A (squares) and L-Phe-CHNA
(circles) on the imprinted (closed points) and blank (open points)
columns eluted with acetonitrile containing variable amounts of
acetic acid.
higher if compared with the blank column, both for
ochratoxin A and L-Phe-CHNA. The increased reten-
tion times of these molecules are the clear evidence
that L-Phe-CHNA is an appropriate mimic molecule
for the mycotoxin, capable of causing a molecular
imprinting effect during the polymerisation process.
The retention behaviour of the imprinted column can
be interpreted by considering that the molecular re-
cognition mechanism is based on hydrogen bond-
ing between the template (or ochratoxin A) and the
carboxyls of the stationary phase. As a consequence of
the competition between the acetic acid and the eluted
molecule for these carboxyls, increasing amounts of
acetic acid in the mobile phase cause a decrease of
393
the retention of the analytes. Also, the polarity of
the mobile phase governs the elution patterns. The
acetonitrile is more polar than chloroform, and as a
consequence both the ochratoxin A and the template
are eluted with reduced retention times compared to
the same elutions performed with chloroform. It is re-
markable that the ochratoxin A and the template are
recognised in a quite similar manner by the imprin-
ted stationary phase, except for the elution performed
in chloroform modied with less than 0.1% of acetic
acid, where the template is recognised much better
than the mycotoxin. The strongest retention of L-
Phe-CHNA in chloroform with 0.010.05% of acetic
acid could be explained considering the presence of
a porogen memory effect. This effect, well known in
literature (Spivak et al., 1997; Yu & Mosbach, 1997),
enhances the retention of the template when eluted by
using the solvent used as porogen during the poly-
merisation process as a mobile phase. The full elution
pattern observed could be interpreted when consider-
ing that the retention of the template is increased by
the porogen memory effect only when the amount of
acetic acid in the mobile phase is almost negligible.
However signicative amounts of acetic acid suppress
this effect as a consequence of the increased competi-
tion for the binding sites between the template and the
acid.
Molecular recognition of template analogs
To obtain information on the non-covalent interac-
tions between the polymer binding sites and the my-
cotoxin, several analogs of L-Phe-CHNA were eluted
using chloroformacetic acid 0.1% (v/v) as a mobile
phase, as in these experimental conditions the reten-
tion factors for the template and the ochratoxin A are
quite similar.
The effect of the presence (or absence) of an
aminoacidic sub-structure on the molecular recog-
nition of the template was studied by eluting the
D-isomer, four other amino acidic derivatives: L-
Ala-CHNA, L-Gly-CHNA, L-Trp-CHNA, and L-Tyr-
CHNA, the 4-chloro-1-hydroxy-2-naphthoylamideand
the 4-chloro-1-hydroxy-2-naphthoic acid. From the
selectivity factors reported in Figure 4, it is clear
that the presence and the molecular structure of the
amino acidic part are essential for the recognition.
The recognition is less effective more the amino acid
is different from L-phenylalanine, practically reduced
by two thirds when the amino acid is absent (amido-
CHNA and CHNA). The presence of an aminoacidic
Figure 4. Polymer selectivity factors measured for ochratoxin A
and L-Phe-CHNA analogs.
structure which hinders more than L-phenylalanine
(L-Tyr-CHNA and L-Trp-CHNA) or with different
spatial orientation (D-Phe-CHNA) also reduced the
recognition, even if did not suppress it.
The effect of the carboxylic part of the amino
acidic sub-structure, able to form hydrogen bonds,
was studied by considering the molecules tyramine-
CHNA, phenethyl-CHNA and L-PheOMe-CHNA
as analogs. The absence (tyramine-CHNA and
phenethyl-CHNA) or the blocking (L-PheOMe-
CHNA) of the carboxylic function does not com-
pletely suppress the recognition but only reduces it.
The effect of the naphthalenic sub-structure was
studied by considering as analogs the molecules L-
Phe-HNA, L-Phe-NA and L-Phe-CSA. The absence of
the phenolic hydroxyl (L-Phe-NA) causes a marked
reduction of the recognition effect, but also the chlor-
ine atom or the naphtalenic sub-structure itself is
able to participate in the molecular recognition effect,
because their absence causes a large decreasing in it.
Considering the whole pattern of the experimental
selectivity factors, it can be seen that only the sim-
ultaneous presence of the carboxyl, of the phenolic
hydroxyl and of certain particular substructures such
as the chlorine atom assures the whole recognition
of the template. This fact implies that the hydrogen
bond between the template and the polymer is fun-
damental, but to assure the full molecular recognition
effect adjunctive sterical factors are necessary.
Conclusions
The use of a synthetic template that mimics the struc-
ture of ochratoxin A acts successfully as a putative
394
imprinting molecule in accordance with the dummy
template polymerisation technique. The polymer ob-
tained binds both the template and ochratoxin A
through hydrogen bond interactions and sterical t
effects in chloroform and acetonitrile, and the mag-
nitude of this interaction could be conditioned by the
use of limited amounts of acetic acid in the mobile
phase. Even if we consider that it should be possible
to rene the structure of the template and the im-
printing process to obtain more selectivity and better
chromatographic properties, we think that the polymer
described here is suitable to attempt to utilise it to set
up a solid phase extraction procedure for ochratoxin A
in real samples. At present, studies are in progress.
References
Airan JW & Shah SV (1942) Interaction of sulfuryl chloride and
naphtol derivatives. J. Univ. Bombay 10: 131134 (Chem. Abs.
37: 339).
Andersson LI, Paprica A & Arvidsson T (1997) A highly selective
solid-phase extraction sorbent for preconcentration of sameridine
made by molecular imprinting. Chromatographia 46(12): 5762.
Matsui J, Fujiwara K & Takeuchi T (2000a) Atrazine-selective
polymers prepared by molecular imprinting of trialkylmelamines
as dummy template species of atrazine. Anal. Chem. 72(8):
18101813
Matsui J, Fujiwara K, Ugata S & Takeuchi T (2000b) Solid-phase
extraction with a dibutylmelamine-imprinted polymer as triazine
herbicide-selective sorbent. J. Chromatogr. A 889(12): 2531.
Nakajima M, Tsubouchi H, Miyabe M & Ueno Y (1997) Survey
of aatoxin B1 and ochratoxin A in commercial green cof-
fee beans by high-performance liquid chromatography linked
with immunoafnity chromatography. Food Agric. Immunol. 9:
7783.
Olsen J, Martin P & Wilson ID (1998) Molecular imprints as sorb-
ents for solid phase extraction: potential and applications. Anal.
Comm. 35: 13H14H.
Poland AE, Schuller PL, Steyn PS & van Egmond HP (1982)
Physicochemical data for some selected mycotoxins. Pure Appl.
Chem. 54: 22202284.
Poland AE, Nesheim S & Friedman L (1992) Ochratoxin A: a
review. Pure Appl. Chem. 64: 10291046.
Rashid BA, Briggs RJ, Hay JN & Stevenson D (1997) Prelimin-
ary evaluation of a molecular imprinted polymer for solid-phase
extraction of tamoxifen. Anal. Comm. 34(10): 303305.
Scott PM & Trucksess MW (1997) Application of immunoafnity
columns to mycotoxin analysis. J.AOAC Int. 80: 941949.
Sellergren B (1999) Polymer- and template-related factors in-
uencing the efciency in molecularly imprinted solid-phase
extractions. Trends Anal. Chem. 18(3): 164174.
Spivak D, Gilmore MA & Shea KJ (1997) Evaluation of binding
and origins of specicity of 9-ethyladenine imprinted polymers.
J. Am. Chem. Soc. 119(19): 43884393.
Stevenson D (1999) Molecular imprinted polymers for solid-phase
extraction. Trends Anal. Chem. 18(3): 154158.
US Dept. Health, National Toxicology Program (2000) IX Re-
ports on Carcinogens (http://ehis.niehs.nih.gov/roc/ninth/rahc/oc
hratoxina.pdf).
Wilkes JG & Sutherland JB (1998) Sample preparation and high-
resolution separation of mycotoxins possessing carboxyl groups.
J. Chromatogr. B 717(12): 135156.
Yu C & Mosbach K (1997) Molecular imprinting utilizing an amide
functional group for hydrogen bonding leading to highly efcient
polymers. J. Org. Chem. 67(12): 40574064.