A

TucoOaAss©

K

1

liitdOaopgnd

0d

Analytica Chimica Acta 584 (2007) 343–349

Simultaneous enzyme immunoassay for the screening ofaflatoxin B1 and ochratoxin A in chili samples

Debjani Saha, Debopam Acharya, Dipika Roy, Dilip Shrestha, Tarun K. Dhar ∗Drug Development, Diagnostics and Biotechnology Division, Indian Institute of Chemical Biology,

4 Raja S.C. Mullick Road, Jadavpur, Calcutta 700032, India

Received 4 September 2006; received in revised form 13 November 2006; accepted 15 November 2006Available online 19 November 2006

bstract

Membrane-based immunoassay has been developed for simultaneous estimation of aflatoxin B1 (AFB1) and ochratoxin A (OA) in chili samples.he combined estimation of both the mycotoxins is more economical in respect of time, work and materials than two separate assays. The methodses a low cost test device consisting of a membrane with immobilized anti-AFB1 and anti-OA antibodies and a filter paper attached to a polyethyleneard below the membrane. It allows direct analysis of sample extracts containing substantial amount (40%) of methanol. This permits the usef two-fold diluted sample extracts resulting in minimum dilution error. The limit of quantitation obtained was 2 and 10 �g kg−1 for AFB1 andA, respectively. The tolerance of 40% methanol was found to be due to the application of small size (0.8 mm diameter) spots on membranes,s the tolerance decreases to 20% with gradual increase in spot size. The combined method is capable of producing acceptable results to analyze

FB1 and OA in chili with accuracy and precision. The AFB1 and OA values obtained for spiked and naturally contaminated chili samples by the

imultaneous method were in good correlation with those measured by individual ELISA. The method offers a simple, rapid and cost-effectivecreening tool to meet the requirements of the rapidly evolving EU legislation.

2006 Elsevier B.V. All rights reserved.

wAfrhchTC[O

eywords: Aflatoxin B1; Ochratoxin A; Simultaneous method

. Introduction

Mycotoxin contamination is one of the most common prob-ems concerning food safety as it causes a variety of toxic effectsn human and animals [1,2]. The alarming feature of mycotoxinss their frequent occurrence in food items in different combina-ions, the consumption of which may exert a greater degree ofamage to health [3–5]. Among different mycotoxins, AFB1 andA are quite common contaminants, which can occur jointly inwide range of food commodities including spices [6,7]. Moldsf the genus Aspergillus and Penicillium are most important inroducing these mycotoxins. AFB1 is a known potent carcino-en in a number of animal species including humans. OA is

ephrotoxic and is implicated in the occurrence of a fatal kidneyisease in humans.

∗ Corresponding author. Tel.: +91 33 2473 5112; fax: +91 33 2473 0284/5197.E-mail address: tkdhar@iicb.res.in (T.K. Dhar).

OgTiO

p

003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved.oi:10.1016/j.aca.2006.11.042

Chili is a popular spice, consumed by many people around theorld. In recent years, several European countries have foundFB1 levels exceeding legal limit in chili samples imported

rom developing countries [8–11]. Further, in a survey of spicesecently commissioned by the Food Standard Agency of UK,igh levels of AFB1 or OA or both toxins were reported inhili samples [12]. The European Scientific Committee for foodas fixed the legal limit for AFB1 at 5 �g kg−1 for spices [13].hough there is currently no legal limit for OA, the Europeanommission has been discussing a limit of 10 �g kg−1 in spices

14]. Several countries have set maximum tolerable levels forA ranging from 1 to 50 �g kg−1 for food.

The current approach for simultaneous detection of AFB1 andA in food samples using HPLC with immunoaffinity chromato-raphic cleanup is laborious, costly and time consuming [15,16].herefore, there has been an increasing demand from the food

ndustry for a simple combined detection method for AFB1 andA to meet the requirements of various regulatory authorities.

We have recently reported an analytical device capable oferforming immunofiltration-based assay for the detection of

3 imica

mfrfea

ticthrcy

2

2

lfbeTLtlH(fruata

2

gfAwbTtatw8wsWcp

Adb

2

i5odae

2

ooO3a

2

mAdbppo25wavbpb(

aoat(

2

44 D. Saha et al. / Analytica Ch

ycotoxins in food samples [17,18]. The method is based onocused absorption of an applied standard (or sample) andeagents through an aqueous network of capillary channelsormed between a membrane and a wetted absorbent body. How-ver, the method requires multiple washing steps due to thepplication of signal amplification steps.

We explored the possibility of using the device for the detec-ion of multiple mycotoxins and succeeded in developing anmproved device for simultaneous detection of AFB1 and OA inhili samples. Recently, few immunological methods for simul-aneous estimation of mycotoxins using biosensors [19–23]ave been developed. However, these methods have limitationsegarding the re-use of the immunosurface and higher reagentonsumption, and lack the ability to perform simultaneous anal-sis of multiple samples [24].

. Experimental

.1. Materials and chemicals

Nitrocellulose membrane, pore size 0.45 �m, was from Mil-ipore Corporation, Bedford, USA. Filter paper (No. 3) wasrom Whatman International Ltd., Maidstone, England. Flat-ottomed polystyrene microtitre plates (Maxisorp) and theight-channel microplate washer were from Nunc, Denmark.he automatic microtitre plate reader (Multiscan MS) was fromabsystems, Finland. Semi-rigid polyethylene sheets, adhesive

ape and analytical grade buffer salts were purchased from theocal market. All toxins, casein, bovine serum albumin (BSA),RP, goat anti-rabbit IgG, Tween 20, 4-chloro-1-naphthol

4CN), 3,3′-diaminobenzidine (DAB) and other chemicals wererom Sigma, St. Louis, USA. Densitometric analysis was car-ied out with an Image Scanner (Amersham Pharmacia Biotech)sing the Magic Scan software (version 4.5) for image scanningnd the Image Master Total lab software (version 1.11) for quan-itation [17]. HPLC analyses of chili samples were carried outs described previously [25].

.2. Reagent preparations

AFB1-O-carboxymethyloxime-BSA and OA-BSA conju-ates were prepared in the laboratory and used as immunogensor raising polyclonal antibodies in rabbits [26,27]. The anti-FB1 antibody has already been characterized by ELISA andas found to be highly specific for AFB1 [17]. The anti-OA anti-ody has been similarly found to be highly specific for OA [28].he AFB1-HRP, OA-HRP and OA-casein conjugates were syn-

hesized by NHS-ester method and purified by extensive dialysisgainst phosphate buffer [27]. The stock solution of the mix-ure of enzyme conjugates (AFB1-HRP + OA-HRP conjugate)as prepared in the assay buffer (Tris–HCl buffer, 50 mM, pH.0, 0.9% NaCl, and thimersal 0.01%) containing 0.25% caseinith 0.01% m-cresol) and stored at 4 ◦C. AFB1 and OA stock

olutions (0.25 mg mL−1 in acetonitrile) were stored at −20 ◦C.orking standards of a mixture of AFB1 and OA in the con-

entrations of 2, 10; 10, 20; 20, 50 and 50, 100 �g kg−1 wererepared in blank chili extracts. In the case of ELISA, individual

wosw

Acta 584 (2007) 343–349

FB1 standards of 2, 5, 10, 25 and 50 �g kg−1, and OA stan-ards of 2, 10, 25, 50, 100 and 500 �g kg−1 were prepared inlank chili extract.

.3. Sample preparation

Chili pods were dried in hot oven (60 ◦C) for 24 h and milledn a coffee grinder; 2 g portions of each sample were soaked inmL methanol–water (80:20) for 15 min [29–31]. After thor-ugh mixing it was centrifuged and the clear supernatant wasecanted. The extracts were diluted two-fold with assay buffernd used directly in simultaneous assay. In the case of ELISA,xtracts were diluted 10-fold in the assay buffer and used.

.4. Preparation of spiked sample

Spiked chili samples in the range of 2–16 and 10–80 �g kg−1

f AFB1 and OA, respectively were prepared by adding 500 �Lf methanol containing appropriate concentrations of AFB1 andA to 2 g of powered chili. The samples were incubated at7 ◦C for 24 h and extracted with aqueous methanol as describedbove.

.5. Preparation of membranes

Briefly, nitrocellulose membrane (56 mm × 70 mm) wasarked with a pencil to give ten 14 mm × 14 mm squares.circle of about 6 mm diameter was marked in the mid-

le of all the squares. The membrane was soaked in blottinguffer (Tris–HCl, 20 mM, pH 8.0 containing 0.9% NaCl),laced in between the adjacent sides of the pre-wetted filteraper (140 mm × 120 mm), and pressed as described previ-usly [28]. Anti-AFB1 and anti-OA antibodies diluted 500- and00-fold with Tris-buffered saline (50 mM, pH 8.0) containing0 �g mL−1 of BSA and 50 �g mL−1 of casein, respectively,ere applied around the circle in a volume of 1 �L each sep-

rately in duplicate using a syringe. The membrane was dried;acant sites blocked with 0.4% casein in carbonate-bicarbonateuffer (50 mM, pH 9.6) and washed (Tris–HCl buffer, 20 mM,H 8.0, containing 2.9% NaCl and 0.05% Tween 20). The mem-rane was again dried and cut with a blade into two strips70 mm × 28 mm each).

A rectangular piece of filter paper (100 mm × 80 mm) wasttached parallel to and about 1 cm away from the shorter edgef a polyethylene card (140 mm × 120 mm) with a tape havingdhesive on both the sides. Then, the unspotted top square ofhe membrane strip was attached over one side of the filter paperFig. 1).

.6. Simultaneous method

The membrane and the filter paper of the device were first

etted with water; after shaking off excess water, it was placedver a horizontal surface and rolled with a rimless test tube. Thetandards or samples, two-fold diluted in assay buffer (25 �L)ere applied at the center of each square of the membrane. Each

D. Saha et al. / Analytica Chimica

Fig. 1. The principle of the simultaneous immunoassay method for AFB1 andOA. A: filter paper; B: membrane strip; C: polyethylene card; : (black) anti-Air

spetcatCstocc(Sw

2

rbowaRTbst

eo

3

3

tppubf[aaablstaosbttd

itzdbfiostftiprit

3

oabfij

FB1, (brown) anti-OA; : (black) AFB1-HRP, (brown) OA-HRP. (Fornterpretation of the references to colour in this figure legend, the reader iseferred to the web version of the article.)

potted zone was washed by adding 2× 25 �L of 50 mM phos-hate buffer saline. The excess water absorbed in the filter wasxtruded by rolling a glass tube over the membrane and gen-ly sponging with a tissue paper. Then a mixture of the enzymeonjugate (AFB1-HRP + OA-HRP) in assay buffer (25 �L) wasdded, washed, and the excess fluid in the filter paper extruded byhe procedure described above. Finally, the substrate solution (4-N + DAB; 2× 25 �L) was added and incubated for 2 min. The

trips were washed under tap water and AFB1 and OA concen-rations in the unknown sample were determined either visuallyr by scanning and analyzing by densitometry. The calibrationurves for AFB1 and OA were constructed by plotting the per-entages of B/B0 values (x-axis) against the toxin concentrationsy-axis) applied. The limit of detection (LOD) was estimated at 2.D. below the zero standard (n = 10). Double of the LOD valueas taken as the limit of quantitation (LOQ).

.7. ELISA procedure

Microtitre plate wells were coated with 150 �L of goat anti-abbit IgG (10 �g mL−1, 18 h at 4 ◦C) and the vacant siteslocked with 0.4% casein. Aliquots of 50 �L of AFB1 standardsr samples (diluted 10-fold in assay buffer) were added to theells followed by 100 �L of a mixture of anti-AFB1 antibody

nd AFB1-HRP conjugate. The plate was incubated for 30 min atT and washed (phosphate buffer, 50 mM, pH 7.6 with 0.05%ween 20); the quantity of bound peroxidase was determinedy adding 150 �L of TMB substrate solution. The reaction wastopped after 5 min by adding 50 �L of 2N sulphuric acid and

he absorbance at 450 nm was measured.

The protocol for OA assay was similar to AFB1 with thexception that immunoincubation and substrate incubation peri-ds were 2 h and 15 min, respectively.

mcbr

Acta 584 (2007) 343–349 345

. Results and discussion

.1. Improved device

The spotting of antigen or antibody on the membranes byhe conventional methods leads to heterogeneous spot mor-hologies and deposition inconsistencies. To circumvent theseroblems, a novel method of spotting on the membranes wassed which is based on focused absorption of an applied anti-ody solution through an aqueous network of capillary channelsormed between the membrane and a wetted absorbent body28]. The antibody solutions were applied over the markedreas of the membrane, and were uniformly absorbed throughn aqueous network of capillary channels without using ofny pump. The spot intensities obtained within as well asetween strips were quantified by densitometry. The calcu-ated CVs of the signal intensity of anti-AFB1 and anti-OApots were 4 and 4.6% (n = 16), respectively. During the assay,he application of 25 �L of standard (or sample) and reagentst the center of spotted zone was sufficient for simultane-us absorption through all the four spots. Although we havepotted only two antibodies in duplicate, the number of anti-odies and spots in each spotted zone may be increased andhe spot size further reduced by spotting lower volumes ofhe antibody solution, but this has not been investigated inetail.

In our previous communication, we have demonstrated themportance of the area of the filter paper and the water con-ent in maintaining uniform flow rates through each spottedone [17,18]. As demonstrated, flow rates and assay repro-ucibility are affected if excess water is removed, as the contactetween the membrane and the filter paper becomes insuf-cient and non-uniform. On the other hand excess wettingf the filter paper reduces the flow rate and allows lateralpreading [17]. To improve the assay reproducibility and prac-icability, an improved device was constructed, which differsrom the reported one in having a filter paper attached belowhe membrane over a polyethylene card (Fig. 1). This mod-fication has the following advantages: (a) the placement ofre-wetted filter paper below the membrane surface was notequired, and (b) optimal water content could be maintainedn the filter paper merely by shaking the device two to threeimes.

.2. Feasibility of simultaneous analysis of AFB1 and OA

The principal advantage of immunoassays is the specificityf detection, which is derived from the selective nature ofntibody–antigen binding. Thus, it is possible to use anti-odies against AFB1 and OA spotted in a single membraneor their simultaneous detection in a single immunoassay. Asllustrated in Fig. 1, we added a mixture of two enzyme con-ugates (AFB1-HRP + OA-HRP) to the spotted zones of the

embrane. Visualization by the addition of substrate showedolour development in both the anti-AFB1 and anti-OA anti-ody spotted zones. However, in the absence of either of theespective enzyme conjugates (AFB1-HRP or OA-HRP), no

346 D. Saha et al. / Analytica Chimica Acta 584 (2007) 343–349

Fig. 2. Simultaneous detection of AFB1 and OA by membrane immunoassay.Buffer standards each containing AFB + OA of value: 0 pg (a), 5 pg (b), 10 pg(s

caaowO

3

auctlcahcatrO

sewdbielra

psot8oOa

Fo

aosm

3

ss1cO1E

3

The influence of methanol concentration in the AFB1 and OAassay was studied by both simultaneous and ELISA methods.In the simultaneous method, the assay buffer containing 20 and

Table 1Cross-reactivity of compounds structurally related to AFB1 and OA determinedby the simultaneous methoda

Compound Cross-reactivityb (%)

Anti-AFB1 antibody Anti-OA antibody

Aflatoxin B1 100 0Aflatoxin B2 13.5 0Aflatoxin G1 11.5 0Aflatoxin G2 0.6 0

1

c), 25 pg (d), 50 pg (e), 100 pg (f), 250 pg (g) and 500 pg (h) were used perpotted zone.

olour development occurs over respective spotted zones. Thessay performed simultaneously using a mixture of standardsnd enzyme conjugates gave results almost identical with thatbtained individually. Therefore, it could be concluded that thereas no cross-reaction during simultaneous analysis of AFB1 andA.

.3. Optimization of the simultaneous method

To determine the optimal dilutions of immunoassay reagents,checkerboard type of assay was performed initially for individ-al analytes using different dilutions of antibodies and enzymeonjugates. It was necessary to ensure that the spot intensi-ies of the buffer zero were high enough to be observable andow enough to be differentiated in presence of a minimal con-entration of analytes. Using appropriate dilutions of antibodynd enzyme conjugate, we optimized the assay sensitivity. Withigh dilution of antibody (1:600), sensitivity increased but theolour intensity decreased. Anti-AFB1 antibody diluted 1:500nd AFB1-HRP conjugate diluted 1:10,000 were found to behe most suitable by trial and error. In the case of OA, the bestesults were obtained with 1:200 anti-OA antibodies and 1:1000A-HRP conjugate dilutions.

We investigated the effect of filtration of the substrate byimultaneous assay. After the addition of zero standard andnzyme conjugates over spotted zones of the membrane, itas visualized by filtration of the substrate. The results pro-uced 60% increase in spot intensity compared to that achievedy pouring the substrate solution over membrane surface andncubating for 2 min [17,25]. It may be pointed out that thisnhancement of spot intensity was achieved by using 50-foldess substrate volume. The interpretation of the higher amplifiedesponse may be ascribed to the rapid accumulation of substratet the site of filtration.

Competitive simultaneous analysis for AFB1 and OA wereerformed using the optimized dilutions. Fig. 2 shows thecanned image of the membrane strip after the assay. The LODsf the method by densitometry were 5 and 25 pg per spot, respec-ively for AFB1 and OA with corresponding B/B0 values of8 ± 2.4 and 96 ± 2.0%, respectively. The dose-response curves

btained for AFB1 and OA are shown in Fig. 3. The AFB1 andA concentrations required to achieve 50% inhibition were 43nd 280 pg per spot.

OO

ig. 3. Comparison of the dose response curves of AFB1 (�) and OA (�)btained by simultaneous method using densitometry.

To determine the specificity of anti-AFB1 and anti-OAntibodies, cross-reactivities were tested with aflatoxin andchratoxin analogues relative to AFB1 and OA. Table 1hows the cross-reactivity that was found by the simultaneousethod.

.4. Optimization of the ELISA

The performances of AFB1 and OA ELISA based on the one-tep double antibody solid-phase (DASP) format using buffertandards were investigated using the same reagents. Dilutions:60,000 for anti-AFB1 antibody and 1:40,000 for AFB1-HRPonjugate were found to be optimum. In the case of OA, anti-A antibody of 1:15,000 and OA-HRP conjugate dilution of:20,000 were found to be optimum. Fig. 4 shows the optimizedLISA standard curves for AFB1 and OA.

.5. Effect of methanol concentration

chratoxin A 0 100chratoxin B 0 10.7

a Assay conditions were the same as described in Section 2.b Cross-reactivity (%) = (IC50 of AFB1 or OA/IC50 of other compound) × 100.

D. Saha et al. / Analytica Chimica Acta 584 (2007) 343–349 347

F(

4sAtnv

oroe

3

bupoaisAc6csrsw4tn

3

b

Fig. 5. Effect of methanol on inhibition of spot intensities using membranes con-tca

se[i9o

rATasT(1mzeeEAt

dppfrsnta

ig. 4. Comparison of the optimized dose response curves of AFB1 (�) and OA�) obtained by individual ELISA using the same reagents.

0% methanol showed no significant differences in spot inten-ities compared to buffer control. The dose–response curves ofFB1 and OA in assay buffer containing 40% methanol with

he those of standards prepared in assay buffer showed no sig-ificant difference in spot intensities, slopes of the curve or I50alues were obtained.

In contrast, the ELISA results showed significant influencef methanol on both AFB1 and OA assay performance due toetardation of colour development. The optimal concentrationf methanol that could be used in buffer without any significantffect was found to be 10%.

.6. Effect of spot size on methanol tolerance

In our previous studies using membranes spotted with anti-ody of 6 mm diameter, methanol in the range of 15–20% wassed in the assay without any effect on spot intensities com-ared to buffer control [17,25]. We reasoned that the tolerancef such a significant concentration of methanol in the presentssay might be due to the difference in the spot size of themmobilized antibody (6 mm versus 0.8 mm). Accordingly, wepotted 1, 5, 10, 15, 20 and 25 �L spotting volumes of anti-FB1 antibody (1:500 dilution) over the membranes to obtain

orresponding spots of diameter approximately 0.8, 2, 3, 4, 5 andmm, respectively. The strips were then assayed under identicalonditions in presence of 20 to 80% methanol and inhibition inpot intensities compared to buffer control was investigated. Theesults (Fig. 5) plotted as percentage inhibition of signal inten-ity against spot size showed the methanol tolerance decreasesith increasing spot size. Maximum tolerance of methanol (up to0%) was observed for spot of 0.8 mm diameter. Further reduc-ion in spot size may increase methanol tolerance, but we haveot investigated this.

.7. Analysis of chili samples

The use of aqueous methanol in various proportions hasecome popular for the extraction of AFB1 or OA from food-

m

cd

aining spots of approximately 0.8, 2, 3, 4, 5 and 6 mm diameter size. Methanoloncentration used: 20% (�), 30% (�), 40% (�), 60% (©) and 80% (�) inssay buffer. AFB1 assay was carried out under identical assay condition.

tuff [32]. We have used 80% methanol in water (v/v) forxtraction of chili samples as it gave better extraction yield8,29–31]. The average recoveries for AFB1 and OA from non-nfected samples obtained by this method were in between3–120 and 90–110%, respectively, within the spiking rangef 10–50 �g kg−1.

To evaluate the accuracy of the method, extracts of natu-ally contaminated samples were spiked with different levels ofFB1 + OA and analysed in duplicate. Results summarized inable 2 show the excellent accuracy of the method. The intra-nd inter-assay variabilities were determined by assaying twoamples containing low and high concentrations of AFB1 + OA.he intra-assay CVs were 6, 8% for AFB1 and 5, 7% for OA

n = 8). The inter-assay CVs were 8, 11% for AFB1 and 8,1.4% for OA (n = 3). The reproducibility of the method deter-ined by assaying non-infected chili samples over all the spotted

ones were excellent (n = 8, CV < 6%). The AFB1 and OA lev-ls of spiked and naturally contaminated chili samples werestimated and compared with values obtained from individualLISA. Regression coefficients between the two methods forFB1 and OA were determined to be 0.98 and 0.99, respec-

ively.The occurrence of AFB1 and OA in chili samples was

etermined by screening 16 commercially available samplesurchased in and around Calcutta from retail stores. Sam-les were reddish-coloured pods and visually appeared freerom any damage or infection from fungi and insects. Theesults by the simultaneous method showed that only twoamples were contaminated with trace amount of AFB1 andone with OA. The AFB1 detected by HPLC in two con-aminated samples were 2.3 and 9.0 �g kg−1, respectively,gainst 1.8 and 8.4 �g kg−1 determined by simultaneous

ethod.Thus, the data on the analytical parameters (Table 3) indi-

ate that the new method for simultaneous AFB1 and OAetection in chili is sufficiently reproducible, accurate, sen-

348 D. Saha et al. / Analytica Chimica Acta 584 (2007) 343–349

Table 2Analytical recovery of AFB1 and OA from naturally contaminated chili extract

Chili Endogenousa (�g kg−1) Added (�g kg−1) Calculated (�g kg−1) Foundb (�g kg−1) Recovery (%)

Sample no. 1AFB1 1.8 ± 0.4 5 6.8 6 ± 1.2 88

40 41.8 40 ± 2.5 9680 81.8 83 ± 2.0 101

OA 5.4 5 10.4 9 ± 2.0 8740 45.4 47 ± 3.5 10480 85.4 86 ± 4.0 101

Sample no. 2AFB1 8.4 ± 1.2 5 13.4 12 ± 1.6 90

40 48.4 46 ± 3.0 9580 88.4 86 ± 2.5 97

OA 1.6 5 6.6 7 ± 3.0 9440 41.6 40 ± 2.5 9680 81.6 80 ± 3.5 98

srtriTawhsf6i

bcmmrtlaf

TAO

A

DICCAA

4

pcadc((msfi

lfoc

a OA was artificially contaminated.b Mean of duplicate determination ± S.D.

itive and specific. Taking into account the sample dilutionequirement and the operative working range, the LOQ of simul-aneous method for AFB1 and OA was 2 and 10 �g kg−1,espectively. For determining OA content below this level, its essential to cleanup the sample as described previously [25].he total time required for wetting of membrane, sequentialddition of standards or samples and enzyme conjugate andashing buffer after each step was about 3 min. The methodas one incubation step of 2 min after the addition of sub-trate solutions. The total assay time required in our laboratoryor analyzing four samples in a single test card was withinmin (sample extraction time and densitometric analysis not

ncluded).The stability of the coated strips and reagents were examined

y periodically assaying standards. The results showed that theoated strips stored at RT (26–30 ◦C) in the dark were stable forore than 3 months. The stock solution of enzyme conjugateixture (AFB1-HRP + OA-HRP at 1:2000 and 1:200 dilutions,

espectively) in presence of 0.01% m-cresol was stable for more◦

han 6 months at 4 C; however without this, 80% activity is

ost within 2 days. When stored at 37 ◦C, it looses 20% of itsctivity within 15 days. Thus, all the reagents were quite stableor long-term use.

able 3nalytical characteristics of the simultaneous method and ELISA for AFB1 andA detection in chili samples

nalytical characteristic Simultaneousmethod

Individual ELISA

AFB1 OA AFB1 OA

etection limit (�g kg−1) 2 10 2 2C50 (�g kg−1) 10 58 10 25V, intra-assay (%) 6–8 5–7 4–6 5–7V, inter-assay (%) 8–11 8–11.4 6–9 7–11ssay time 6 min for combined assay 35 min 2 h 15 minverage recovery (%) 92–120 93–110 95–120 90–105

A

D

R

. Conclusions

The potential of a simple and highly cost-effective device forerforming simultaneous rapid detection of AFB1 and OA inhili samples in a single assay has been demonstrated. The maindvantages of the method are five-fold: (i) it is more rapid; (ii)oes not require costly, laborious and time consuming sampleleanup and enrichment steps, allowing considerable savings;iii) the consumption of washing buffer and substrate solution50-fold) is markedly reduced; (iv) the sample dilution error isinimum as the extract is only two-fold diluted; (v) is highly

uitable for initial cost-effective screening of chili samples undereld conditions.

Although we have limited our investigation to two ana-ytes, it is obvious that this simple approach may be utilizedor simultaneous detection of multiple analytes. The possibilityf expanding this technique to other types of food matrices isurrently under investigation.

cknowledgements

We thank the Council of Scientific and Industrial Research,elhi, for Research Fellowships to DS and DA.

eferences

[1] W.J. Bennett, M. Klich, Clin. Microbiol. Rev. 16 (2003) 497.[2] H.S. Hussein, J.M. Brasel, Toxicology 167 (2001) 101.[3] G.J.A. Speijers, M.H.M. Speijers, Toxicol. Lett. 153 (2004) 91.[4] M. Sedmikova, H. Reisnerova, Z. Dufkova, I. Barta, F. Jilek, Vet. Med.

Czech. 46 (2001) 169.[5] P.B. Wangikar, P. Dwivedi, N. Sinha, A.K. Sharma, A.G. Telang, Toxicol-

ogy 215 (2005) 37.

[6] J.W. Park, E.K. Kim, D.H. Shon, Y.B. Kim, Food Addit. Contam. 19 (2002)

1073.[7] Q.A. Mandeel, Mycopathologia 139 (2005) 291.[8] M.L. Martins, H.M. Martina, F. Bernardo, Food Addit. Contam. 18 (2001)

315.

imica

[

[[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

D. Saha et al. / Analytica Ch

[9] B. Fazekas, A. Tar, M. Kovacs, Food Addit. Contam. 22 (2005) 856.10] R. Russell, M. Paterson, Food Control, doi:10.1016/j.foodcont.2006.

04.005 (available online April 29, 2006).11] K. Jorgensen, Food Addit. Contam. Suppl. 1 (2005) 26.12] Food Standard Agency, UK, Survey of Spices for Aflatoxins and Ochra-

toxin A. Food Survey Information Sheet, 73/05, 2005.13] European Commission, Commission regulation 472/2002 of March 12,

2002 amending regulation (EC) No. 466/201 setting maximum levels forcertain contaminants in foodstuffs, Off. J. Eur. Commun. L75 (2002) 18–20.

14] European Commission, Commission regulation 123/2005 of January 26,2005 amending regulation (EC) No. 466/2001 as regards ochratoxin A,Off. J. Eur. Commun. L25 (2005) 3–5.

15] R.J. Garcia-Villanova, C. Cordon, A.M. Gonzalez Paramas, P. Aparicio,M.E. Garcia Rosales, J. Agric. Food Chem. 62 (2004) 7235.

16] D. Chan, S.J. MacDonald, V. Boughtflower, P. Brereton, J. Chromatogr. A1059 (2004) 13.

17] A. Pal, T.K. Dhar, Anal. Chem. 76 (2004) 98.18] A. Pal, D. Acharya, D. Saha, D. Roy, T.K. Dhar, Anal. Chem. 76 (2004)

4237.19] E. Schneider, E. Usleber, E. Martlbauer, J. Agric. Food Chem. 43 (1995)

2548.20] B. van der Gaag, S. Spath, H. Dietrich, E. Shigter, G. Boonzaaijer, T. Van

Osenbruggen, K. Koopal, Food Control 14 (2003) 251.21] E. Schneider, V. Curtui, C. Seidler, R. Dietrich, E. Usleber, E. Martlbauer,

Toxicol. Lett. 153 (2004) 113.

[

[

Acta 584 (2007) 343–349 349

22] K.E. Sapsford, M.M. Ngundi, M.H. Moore, M.E. Lassman, L.C.Shriver-Lake, C.R. Taitt, F.S. Ligler, Sens. Actuators B 113 (2006)599.

23] N. Adanyi, I.A. Levkovets, S. Rodriguez-Gil, A. Ronald, M. Varadi, I.Szendro, Biosens. Bioelectron. doi:10.1016/j.bios.2006.02.015 (availableonline April 5, 2006).

24] M.A. Gonzalez-Martinez, R. Puchades, A. Maquieira, Trends Anal. Chem.18 (1999) 204.

25] A. Pal, D. Acharya, D. Saha, D. Roy, T.K. Dhar, J. Food Prot. 68 (2005)2169.

26] D. Bhattacharya, R. Bhattacharya, T.K. Dhar, J. Immunol. Meth. 230 (1999)71.

27] J. Das Sharma, C. Duttagupta, E. Ali, T.K. Dhar, J. Immunol. Meth. 184(1995) 1.

28] D. Saha, D. Acharya, T.K. Dhar, Anal. Bioanal. Chem. 385 (2006)847.

29] K. Thirumala-Devi, A.A. Maya, G. Reddy, S.V. Reddy, P. Delfosse, D.V.R.Reddy, J. Agric. Food Chem. 48 (2000) 5079.

30] S.V. Reddy, D.K. Mayi, M.U. Reddy, K. Thirumala-Devi, D.V.R. Reddy,Food Addit. Contam. 18 (2001) 553.

31] M.M. Ngundi, L.C. Shriver-Lake, M.H. Moore, M.E. Lassman, F.S. Ligler,C.R. Taitt, Anal. Chem. 77 (2005) 148.

32] M.P. Scott, in: J.W. DeVries, M.W. Trucksess (Eds.), Mycotoxins and FoodSafety, Kluwer Academic Publishers, Dordrecht, The Netherlands, 2002,pp. 117–134.