Mycotoxin Research, Vol. 4 (1988)

A monoclonal antibody-based enzyme-linked immunosorbent assayof aflatoxin 8, in peanut products

o Kawamura' , S Nagayama', S Sato' , K Ohtani" , I Uen02 , and Y Ueno' *1 Department of Toxicology and Microbial Chemistry,

Faculty of Pharmaceutical Sciences, Science University of Tokyo,Ichigaya, Tokyo 162, Japan

2 The Institute of Medical Sciences,University of Tokyo,Shiroganedai, Tokyo 108, Japan

* Off print request to: Y Ueno at the address 1 above

Abstract

An improved enzyme-linked immunosorbent assay (ELISA) combined with mono­clonal antibody (MAb) and one-step extraction method was established for theestimation of aflatoxin B, (AFB,) in a peanut product. AFB, was converted toAFB,-oxime, and then conjugated with bovine serum albumin (BSA). Spleen cellsfrom mice immunized with AFB,-BSA conjugates were fused with myeloma cells.After double selection with AFB,-ovalbumin (OVA) and carbodiimide-modifiedOVA, five stable hybridoma cells secreting anti-AFB, MAbs (AF 1, AF 2, AF 3,AF 4, and AF 5) were cloned. Using these anti-AFB, MAbs, we developed theindirect competitive ELISA (cELlSA) with alkaline phosphatase (ALP) - labeledsheep anti - mouse IgG as marker and the direct cE LISA with AFB,-oxime horse­radish peroxidase (POD) as marker.

The minimum detectable limits of the indirect cELlSA with AF 1,2,3,4, and5 were 5, 5, 5, 5, and 50 pg of standard AFB, per assay, respectively, and those ofthe direct cELlSA with AF 1, 3, 4, and 5 were 2.5,5,25, and 100 pg of standardAFB, per assay, respectively. The cross reactivity of each toxins with these MAbsin the indirect cELlSA was as follows: (a) AF 1 and AF 2 were reactive with AFB2as well as AFB" weakly with AFG2 > AFG, > aflatoxicolll (COL II) > aflatoxicoll(COL J) and less weakly with other aflatoxins; (b) AF 3 and AF 4 were reactivewith COL II as well as AFB" weakly with COLI> AFO, and less weakly withothers; (e) AF 5 was AFO, as well as AFB" weakly with COLli> AFG2 > COL Iand less weakly with others.

The 60% aqueous methanol extracts of oil-roasted blanched peanuts ("butterpeanut"), naturally contaminated with AFB" were assayed by the direct cELlSAwithout further purification. The direct cELlSA with the most sensitive MAb AF 1was allowed to determine 1 ng of AFB, per g samples.

Introduction

AFB1 , a potent mutagen and hepato-carcinogen produced by fungi Aspergillusflavus and A. parasiticus, is often detected in agricultural commodities, feed, andfood (1, 2). Epidemiological surveys have revealed a close association between ahigh incidence of primary liver cancer (PLC) and a high level of AFB1 in food in

75

endemic areas such as South Africa (3), China (4), and Philippines (5). A statisticallynegative correlation between the incidence of PLC and the intake of AFB1 wasreported in the USA (6). In addition, the World Health Organization (WHO) reportedthat hepatitis B virus (HBV) is second only to tobacco (cigarette) products amongthe known carcinogens (7). These approaches lead to present a multietiology theoryon the cause of PLC (4, 8).

In order to solve such problems as the association of AFB1 with PLC, two majorapproaches are now in progress; one is molecular biology. Wogan's and the author'sgroups have demonstrated the over-expression of proto-oncogenes in rodent liversafter AFB1 exposure (9, 10, 11). The other is to develop sensitive and specific metho­dology to estimate the contents of AFB[ in food and AFB1 -modified DNA adductsin the human body. Along this line, several ELISA with antisera (12, 13, 14, 15, 16)and MAbs (17, 18, 19) against AFB1 and DNA adducts have been proposed.

In this report, we describe the production of a specific MAbs against AFBjusing hybridoma technology, and characterize their specificity. An established directcELISA combined with a one-step extraction procedure was applied to estimatethe content of AFBj in oil- roasted blanched peanuts, butter peanuts.

Materials and Methods

Chemicals

Bovine serum albumin (BSA), chicken egg albumin (ovalbumin, OVA), l-ethyl-3,3 - dimethylaminopropyl- carbodiimide (EDPC), POD (type VI), insulin from bovinepancrease, AFBj, AFB2 , AFBza, AFG1 , AFGz, AFM1 , AFQb AFP1 , COLI (naturalisomer [1S] -aflatoxicol) and COL II (natural isomer [I R] -aflatoxicol), were purchas­ed from Sigma Chemical Co (St Louis, MO). Polyoxyethylene, sorbitane monolaurate(Tween 20), p-nitrophenyl phosphate (PNPP), o-phenylenediamine, 2,3,1O,14-tetra­methyl pentadecane (Pristane), and diethanolamine were purchased from WakoPure Chemical Ind (Tokyo). ALP from calf intestine (> 2500 U/ mg) was obtainedfrom Boehringer Mannheim GmbH Biochemicals (Mannheim). Affinity purifiedF(ab)z' fragment sheep anti -mouse IgG (heavy & light chain specific) was purchasedfrom Cappel Lab (West Chester, PA). Polyethylene glycol 4000 was from E Merck(Darmstadt).

Animals and cells

BALB/ c female mice, aged 6 - 8 weeks, were purchased from Sankyo Lab Service(Sizuoka). SP2/0-AGI4 (SP2) and P3-X63-Ag8-653 (6.5.3) mouse myeloma celllines were obtained from Dr J Chiba, National Institute of Health (Tokyo).

AFB," protein conjugates and immunization

AFBj was converted into AFBj-oxime, then conjugated with proteins (12). BALB/cmice were sc injected with 0.02 mg of AFB1 -BSA conjugates emulsified in Freund'scomplete adjuvand (Difco Lab, Detroit), twice with two week interval. Ten dayslater, the third injection of the same amount of conjugates in PBS (50 mM phosphatebuffered saline, pH 7.4) was ip carried out as a booster. After 72 hrs, the spleencells were used for a following cell fusion.

Preparation of anti-AFB, MAbs

The method of Kohler and Milstein (20) modified by Oi and Herzenburg (21) wasused for the fusion as follows. The spleen cells isolated from the immunized mice

76

Mycotoxin Research, Vol. 4 (1988)

were washed 3 times (x 3) with minimum essential medium. The suspensions of3.5x108 spleen cells and 1.17 X 108 or 4.38 X 108 SP2 myeloma cells or 4.38 X 108

6.5.3 myeloma cells were mixed and centrifuged at 1,400rpm for 6 min. The pelletwas loosened by gentle tapping, and 1mL of 50 % polyethylene glycol 4000 inRPMI 1640 (Nissui Pharmaceuticals, Tokyo) containing 20 mM HEPES (Gibco,New York) pre-warmed to 37°C was added to the cells in drop over a period ofone minute with vigorous stirring. The suspension of cells was diluted by an additionof 9mL of RPMI 1640 containing 20mM HEPES (1.8mL/min with vigorous stirr­ing). After HAT selection, the supernatants of hybridoma cells were assayed by theindirect non-competitive ELISA (neELISA) as described below. The hybridomacells from positive wells were propagated in 24 -well plates. The positive cultureswere cloned 5 to 8 days after further growth by a limiting dilution method (21)into 96-well microplates in RPMI 1640 medium which was containing 30% fetalbovine serum (MA Bioproduct, Wallersuille), 0.1 mM hypoxanthine, 161lM thymi­dine, 0.03 % L - glutamine and thymocytes (106 per well) as a feeder layer. Wellswere examined for single clones and rescreened for specific antibody activity 7to 10 days later. Clones secreting anti -AFB1 antibody were grown in HT mediumalone and maintained at a cell density between 105 and 106 cells/mL.

Cloned hybridoma cells in PBS were ip injected into 8 to 10 weeks old BALB/cmice (lx107 cells/A) pretreated by ip injection of 1mL of Pristane, and grew asascites tumors. Immunoglobulin fraction was prepared from the' ascitic fluid by33.3 % saturated ammonium sulfate precipitation. Volumes of IgG fraction wereadjusted to initial volumes of the ascitic fluid used for ammonium sulfate precipita­tion. The concentrations of MAbs were determined by estimating the absorbanceat 280 nm. These MAbs were stored at -70°C.

Isotypes of anti-Al-B, MAbswere determined by Ouchterlony diffusion method.

ELISA procedures

a) Indirect non-competitive ELISA100llL of AFB1-OVA or EDPC-modified OVA (2.5~/mL PBS) were added toeach well of a 96-well microtiter plate (Nunc Immunoplate II, Inter Med Co, Ros­kilde), and incubated at 4°C overnight. The plate was washed x 3 with 0.4 mL PBS/Tween (PBS containing 0.05 % Tween 20) using an automatic washer (ImmunoWasher NK-300, Japan Inter Med Co, Tokyo). The wells were then filled (0.12 mLeach) with 0.1 % OVA in PBS and incubated at 37°C for 1hr to block the unoccupiedsites on the plastic. After incubation, the wells were washed x 3 with PBS/ Tweenand filled with 0.1 mL of culture supernatants of the cultured hybridoma cells.After 1hr at room temperature, the wells were washed x 4 with PBS / Tween, andfilled with 0.1 mL ALP - labeled anti - mouse IgG diluted x 100 with PBS -Tween.After incubation for 1hr at room temperature, the wells were washed x 5 as above,followed by addition of 0.1 mL of PNPP (1 mg/ mL 1M diethanolamine buffer,pH 9.8, containing 0.2 % NaN3 and 0.1 % MgCh). After incubation at 37°C for30 min, the reaction was terminated by addition of 0.03 mL of 3 M NaOH to eachwell. The absorbance at 405 nm was determined in an ELISA reader (MicroplatePhotometer MR-22, Corona Electrics Co, Ibaraki). This indirect ncELISA wasused for the screening of anti-Al-B, MAbs-producing hybridoma cells.

b) Indirect competitive ELISA100ilL of AFB1-BSA (250 ng/mL PBS) were added to each well and incubatedat 4°C overnight. The wells were washed x 3 with PBS/ Tween and each filled with0.12mL of OVA in PBS. After incubation at 37°C for 1hr followed by washingas described above, 0.05 mL of AFB1 standards or related compounds were added,followed by each anti - AFB1 MAb (AF 1, 0.2Ilg/ mL; AF 2, 10 Ilg/mL; AF 3,

77

0.5 p.g/mL; AF 4, 0.5 p.g/mL; AF 5, 0.2.ug/mL in PBS / Tween, which gave 50 %binding of maximum binding). After incubation at 4°C overnight, the plates werewashed x 5 with PBS / Tween. The follow-up procedures were the same as thosefor indirect ncELISA as cited above. The %binding was calculated as follows:

% binding = absorbance of AFB, standards or other compounds / absorbance ofbuffer alone x 100.

This indirect cELISA was used for the characterization of AFB,-MAbs and thepreparation of standard curves.

c) Direct competitive ELISA100,uL of each anti - AFB, MAbs (AF 1, 20 p.g/mL; AF 3, 5 p.g/mL; AF 4, 5 p.g/mL;AF 5, 10p.g/mL in PBS) were added to each well, and incubated at 4°C overnight.The wells were washed x 3 with PBS / Tween, and each filled with 0.12 mL of OVAin PBS. After incubation at 37°C for 1hr, followed by washing as described above,0.05 mL of standard aflatoxins or samples and 0.05 mL of AFB, -POD conjugates(22) (20 p.g/mL in PBS / Tween) was added to each well. After stirring the plateon a microtiter mixer for about 30 sec, the plate was incubated at room temperaturefor 30 min, washed x 5 with PBS / Tween, and the wells were each filled with 0.1 mLof substrate solution (0.4 mg/mL of a-phenylenediamine in 50mM citrate - phos­phate buffer, pH 5.5, containing 2 mM H20 2 ) . The plate was incubated at 37°Cfor 30 min, and the reaction was terminated by addition of 0.1 mL of 2 M H2S0 4 ,

The absorbance at 492 nm was determined on an ELISA reader.

d) Sandwich type ELISAA sandwich type ELISA was utilized to determine the concentrations of the MAbsin culture supernatants. 50,uL of sheep anti - mouse IgG (l,ug/mL in PBS) wasadded to 96 - well microplates. The plates were incubated at 4°C overnight, washedx 3 with PBS / Tween and added 0.1 mL of 0.1 % OVA in PBS. After incubationat 37°C for 1hr, the plates were washed x 4 with PBS / Tween. To each well wereadded 0.05 mL of the culture supernatants or normal mouse IgG (0.1 to 100 p.g/mLin HT medium) and incubated at room temperature for 45 min. The plates werewashed x 4 with PBS/Tween. The following steps were the same as those for indirectncELISA as cited above.

Sample preparation

Oil-roasted blanched peanuts (so-called "butter peanuts"), naturally contaminatedwith 200 ng/g of AFB1 were kindly provided by Dr T Tanaka, Public Health Instituteof Kobe City, Kobe. This sample was mixed in steps with non-contaminated butterpeanuts to yield the final concentrations of I, 5, 10, 20, and 50 ng AFBdg. 5 mLof 60 % methanol in water were added to the 1g samples, and the whole was shakenfor 1min. After centrifugation at 3,000 rpm for 3 min, the supernatants were filteredon a filter paper (Toyo Rosi N02, Toyo Rosi Co, Tokyo), and ImL of the filtratewas diluted with the same volume of PBS /Tween. This extract solution was assayedby direct cELISA.

To examine the influence of crude extract on standard curve for AFB, tests,1.8mL of the extract solution prepared from non-contaminated butter peanutswere spiked with 0.2 mL of AFB, (l.ug/mL in methanol), and this mixture (100 ngAFB} /mL) was diluted with the extract solution prepared from non-contaminatedbutter peanuts at levels of equivalent to contaminated 0.2 to 50 AFB1 pg per ginbutter peanuts.

78

Mycotoxin Research, Vol. 4 ('988)

Results and Discussion

Preparations of AFB,-protein conjugates and hybridoma cells

The molar ratios of AFB] bound with BSA and OVA, as determined by UV-analysis,were 2.7 and 0.1, respectively.

The cell fusions were performed under the following ratios of spleen cells tomyeloma cells: 3 to I (SP2, l st), 8 to I (SP2, 2nd), and 8 to 1 (6.5.3). After fusion,cells were distributed into 704, 688, and 704 wells, respectively. After HAT selec­tion, hybridoma cells were found in 99 %, 98 %, and 68 % of the wells, respectively(Table 1). Indirect ncELISA with AFB]-OVA coated plates revealed that 169,99,and 9 wells were markedly (A40S > 1.0) positive, and 105, 74, and 1 wells wereweakly (A40 S > 0.2) positive, respectively. Culture supernatants also strongly reactedwith EDPC- treated OVA were counted 47, 24, and 1, respectively. This means that122 (74%) out of the total 277 culture supernatants were reactive with EDPC­treated OVA.

H was reported that carbodiimide, often used for the coupling of haptens toproteins, gave rise to N-protein-substituted urea (23). When animals were immunizedwith such protein-hapten conjugates prepared by carbodiimide method, anti - N ­substituted urea antibodies were present in the serum, along with anti - hapten anti­bodies. In the experiments presented here, we revealed that more than 1/4 of thetotal hybridoma cells produced antibodies reactive against N-substituted urea.

In order to select the hybridoma cells secreting only MAbs reactive with AFB],we have introduced a plate coated with EDPC-treated OVA, and consequently,5 stable hybridoma cells secreting anti-Al-B, MAbs were cloned, designated as AF 1,AF 2, AF 3, AF 4, and AF 5. Isotype of AF 5 was IgG j kappa and those of otherswere IgG] lambda. The concentrations of the MAbs (AF 1, 2, 3, 4, and 5) were1.03, 0.63, 2.3, 1.8, and 1.2 ,ug/106 cells/mL/24 hrs in the culture media by thesandwich type ELISA and 5.1, 1.5,8.7,7.2, and 8.7 mg/mL in the ascitic fluids byUV-analysis, respectively.

Tab 1: Preparation of anti·AFB, MAb secreting hybridomas cell lines

Myeloma cell lines SP2/0a SP2/0 6.5.3b TotalSpleencells/myeloma cells 3.0 8.0 8.0No of culture wells 704 688 704 2096Growth positive wells 698 (99%) 678 (98%) 480 (68%) 1856 (89%)

Binding assay forAFB"OVA ++c 169 (24%) 99 (15%) 9( 2%) 277 (15%)

+d 105 (15%) 74 (11 %) 1 ( 0%) 180 (10%)Total 274 (39%) 173(26%) 10 ( 2%) 457 (25%)

Binding assay forAFB,-OVA & EDPC-OVA

AFB,-OVA only 47/169 24/99 1/9 27/277 (26 %)AFB,-OVA> EDPC-OVA 21/169 21/99 4/9 46/277(17%)AFB,-OVA'" EDPC-OVA 101/169 45/99 4/9 159/277 (57 %)

StabIe hybridom a/3/5 1/4 1/4 5/13 (38%)cloned wells

a SP2/0,SP2/0-Ag14b 6.5.3, P3- X63-Ag8-653c++, markedly positive (A40 5 >1.0) in the neELISAd +, weakly positive (A40 5 >0.2) in the neELISA

79

Tab

2:T

hecross

reactivityof

MA

bsin

the

indirectcE

LISA

Afla

toxin

sA

nti-a

flato

xin8,

mo

no

clon

al

antibodiesA

F1

AF

2A

F3

AF

4A

F5

81

27.5"(100.0)b

26.5(100.0)

35.5(100.0)

76(100.0)

370(100.0)

82

20.0(138.0)

20.0(133.0)

2,3

00

.0(

1.4)3.350

(2

3)

>5,000

c-7.4)

G,

190.0(

10.5)198.0

(13.4)

3,0

50

.0(

1.2)2,450

(3.4)

>5,000

c-7.4)

G2

165.0(

12.1)180.0

(14.7)

2,6

00

.0(

2.4)3,150

(2.4)

2,200(

16.8)C

OL

Ic800.0

(3.4)

1,300.0(

2.0)285.0

(13.7)

475(

14.1)3,000

(12.3)

CO

Ll1

d490.0

(5.6)

485.0(

5.5)30.0

(125.0)34

(122.0)815

(45.4)

M,

2,7

50

.0(

1.0)2

,85

0.0

(0.9)

2,4

00

.0(

1.7)1.475

(4.5)

>5,000

(>7.4)

Q,

>5

,00

0.0

c-0.6)

>5

,00

0.0c-

0.5)5

24

.0(

7.2)620

(10.8)

393(

94.1)

P,1

,90

0.0

(1.5)

2,8

00

.0(

0.9)>

5,0

00

.0c-

0.8)>

5,000c-

1.3)>

5,000c-

7.4)

82

">

5,0

00

.0c-

0.6)>

5.0

00

.0c-0.5)

>5

,00

0.0

c-0.8)

>5,000

c-1.3)

>5,000

t>7.4)

eA

mo

un

t(pg)

perassay

yield

ing

50%in

hib

ition

ofthe

ma

ximu

menzym

ea

ctivitYb

Num

bersin

parenthesisin

dica

te%

ofcross

rea

ctivityc

CO

LI,

naturalisom

er[t

Sj-a

ftatcxlco

ld

CO

L11,

naturalisom

er[1

R!·a

flato

xicol

Tab

3:T

hecross

reactivityof

MA

bsin

the

directcE

LlS

A

Afla

toxin

sA

nti-a

flato

xin8,

mo

no

clon

al

antibodiesA

F1

AF

3A

F4

AF

5

8,31.3"

(100.01 b51.5

{100.0}78

(100.0)965

(100.0)

82

65.0(

48.2)2

,95

0.0

(1.7)

4,2

00

(1.9)

>5.000

c-19.3)

G,

81.0(

38.6)8

40

.0(

6.1)965

(8.1)

>5,000

c-19.3)

Gz

84.0(

37.3)1,200.0

(4.3\

810(

9.6\>

5,000c-

19.3)

CO

LIe

4,610.0(

0.7)156.0

(33.3)

80(

97.5)2,100

(46.0)

CO

Llid

3,4

80

.0(

0.9)46.0

(112.0)39

(200.0)2,500

(38.61

M,

507.0(

6.2)690.0

(7.5\

815(

9.6\>

5,000c-

19.3)

Q1

>5

,00

0.0

c-0.6)

>5

,00

0.0c-

1.0)>

5,000(>

1.6)1,240

(63.3)

P,1,140.0

(2.7)

1,120.0(

4.6)850

(9.2)

>5,000

(>19.31

82

">

5,0

00

.0c-

0.6)>

5,0

00

.0c-

1.0)3,500

(2.2)

>5,000

c-19.3\

eA

mo

un

t(pg)

perassay

yield

ing

50%in

hib

iton

ofthe

ma

ximu

menzym

ea

ctivityb

Num

bersin

parenthesisin

dica

te%

ofcross

rea

ctivityC

CQ

LI,

naturalisom

er[1

S]-a

flato

xicol

dC

OL

11;naturalisom

er[1

R]-a

flato

xicol

oCXl

MycOToxin Research, Vol. 4 (1988)

100

......•0--enc"0C.-

CO

50

oa 0.5 2.5' 5 10 25 50 100 250500

AFB 1 (pg/assay)5000

Fig 1: The sensitivities of the indirect cELISA with antl-Al-B, MAbs.The indirect cELISA was carried out with each MAb (IgG fraction) at a concentra­tion to give 50 % of maximum binding in the indirect ncELlSA. The concentrationof AF 1 (0), AF 2 (.), AF 3 (6), AF 4 (A), and AF 5 (0) were 0.2, 10,0.5,0.5, and0.2 pg/m L, respectively. Binding (%) = A;IAa x 100. A was the absorbance at 405 nmin the presence (Ai) or absence (Ao ) of AFB1 .

Sensitivity and cross reactivity of anti-AFB, MAbs

The sensitivity and cross reactivity of 5 MAbs were examined by indirect cELISA,as shown in Fig I and Table I, respectively. The minimum detectable limit of AFB)was 50 pgj assay with AF 5, and this limit was reduced to 5 pgjassay with the otherMAbs. The detection limits were expressed as concentrations which gave significant(over 10%) inhibition. These 5 MAbs were grouped into 3 types according to theircross reactivity to AFB) analogues: (a) AF I and AF 2, reactive to AFB2 as well asAFB); (b) AF 3 and AF 4, reactive to COLlI as well as AFB); and (c) AF 5, reactiveto AFQI as well as AFB).

By the direct cELISA, the sensitivity and cross reactivity of 4 out of 5 MAbswere shown in Fig2 and Table 3. The detectable limits of AFB) with AF 1,3,4,and AF 5 were 2.5, 5, 25, and 100pgjassay, respectively. These results indicatedthat the combination of AF I antibody with the direct cELISA was most sensitive.The cross reactivity in the direct cELISA was similar to the indirect cELISA. Butthis combination was more cross reactive with AFG] (38.6%) and AFG 2 (37.6%)than those of the indirect cELISA. The difference of these cross reactivities wouldbe caused by the inequality in reactivity of the MAbs with AFB)-POD and AFB)­BSA. Therefore. the combination of AF I, and the direct cELISA is the most sen­sitive, but is not applicable for the selective estimation of AFB•.

Our ELiSAs with anti-Af'B, MAbs were compared to previously reported ELiSAswith MAbs. Our ELiSAs with AF I or AF 3 are 2 to 4 times more sensitive than thedirect ELISA using antibody-POD conjugate reported by Candlish et al (24), though

81

100

.......;.:-en 50c

"CC(j)

a

a 0.51

Fig 2: The sensitivities of the direct cELISA with anti-oAFB, MAbs.The direct cELISA was carried out with each MAb (lgG fraction) at which theconcentration of AF 1 (0), AF 3 (6), AF 4 (oA), and AF 5 (0) were 20, 5, 5, and10 pg/mL, respectively. Binding (%) = A;lAo x 100. A was the absorbance at 492 nmin the presence (Ai) or absence (AD) of AF B,.

the cross reactivity is similar to our ELISA with our AF 1 or 3. The indirect ELISAusing ALP-labeled anti-mouse Ig were reported by Dradsted et al (25). Their ELISAis reactive with AFB2a as well as AFB I and much cross-reactive with AFGI , AFG 2 ,

AFG2 a , and 8,9-dihydro-8-(N7-guanyl)-9-hydroxyaflatoxin BI .

Interference of ELISA by methanol and sample matrix

Several papers reported 55 to 70 % methanol as an excellent extraction solventfor AFB I in agricultural products and biological samples (26, 27). However, metha­nol over 60 % methanol gives a great influence on the enzyme activity and bindingcapacity of antibody in our ELISA by modifying physical nature of plates and theassociation between antibodies and antigens. When 60 to 100 % methanol was addedto wells both in indirect and the direct cELISA, the absorbance decreased to about40 to 50 % of the maximum (data not shown). As shown in Fig 2, in the indirectcELISA, the standard curves of AFB] shifted to the right in proportion to theconcentrations of methanol added (Fig 2). Consequently the 50 % maximal bindingswere varied to 28, 73, and 1940 pg/ assay in the presence of 10, 30, and 60 % metha­nol, respectively. In the direct cELISA, as already reported by El- Nakib et al (14),the standard curves in the presence of 10 and 30 % methanol were similar to eachother, and we also observed that the effect of 60% methanol was much less thanin the case of indirect cELISA (Fig 3). Actually. the 50% maximal binding was 31,35, and 60 pg/ assay in the presence of 10, 30, and 60 % methanol, and the detectablelimit was 5 pg/assay in all cases. However, 60% methanol caused about 50% inhibi­tion of enzyme activity. These results indicated that the effect of methanol was

82

Mycotoxin Research, Vol. 4 (1988)

'"'•~......,

50OJC"0C

CD

a

o 10 25 50 100 25010002000(pgl assay)

Fig 3: Effect of methanol on the indirect cE LISA.50pL of AFB1 (0 to 40ngfmL) in 10% (0),30% (e), and 60% (6) methanol-waterwere added to wells coated with AFB1- BSA conjugates (250 ngf m L), and 50 plof anti-Af B, MAb AF 1 (200 ngfmL) were added and mixed. After incubationat 4°C overnight, the wells were washed and reacted with ALP-labeled anti-mousets. and bound enzyme was determined. The CV (%) of inter-assay (n = 6) at 5,50, and 50pgfassay were 6.54,17.2, and 24.2 in 10% methanol, 3.09, 4.57, and16.3 in 30 % methanol, and 4.91, 7.66, and 12.4 in 60 % methanol, respectively.

much less in the direct cELISA than the indirect cELISA. The former could beperformed without marked changes in the specificity and stability of anti-Al-B,MAb if the concentration of methanol in the assay system is below 30 %.

Ram and Hart (15) reported that the mixture ofmethanol:water:N,N-dimethyl­formamide (70 : 29 : 1) did not give great effect on the direct cELISA. However,the enzyme activity in our direct cELISA was inhibited by this solvent to about70% of the control (PBS / Tween alone) (data not presented). Probably, differentconditions in regards to concentrations of antibody and AFB j - POD conjugate,reaction time and temperature, may result different influences even when the samesolvent was used in the ELISAs.

Mass surveys of agricultural products and food are one of gread advantages ofELISA as an alternative tool. Therefore, the introduction of a simplified procedurefor the extraction of suspected contaminants promotes its application in practice.However, it is essential to examine the possible interference from sample matrixbefore the actual application of ELISA. First, we tested what was the concentrationof the non-contaminated butter peanut extract (mg/well) that caused the inter­ference on the direct cELISA. When the extract equivalent to 10mg of non-con­taminated one per well was added, about 12%of inhibition appeared. The significantinterference was not observed when 1:8 diluted with PBS/Tween extract (equivalentto 1.25mg of butter peanut / well) was added (data not shown). Our direct cELISAwith AF 1 was allowed 2.5 pg of AFBj per assay. Therefore, if we select the concen-

83

100

,-.............. 5001Ciicen

a

2.5 5AF8 1

o 10 25 50 100(pg/assay)

250 1000 2000

Fig 4: Effect of methanol on the direct cELISA.AFB, (0 to 40ng/niU in 10% (0), 30% (e). and 60% (A) methanol· water wasadded to wells coated with anti-AFB, MAb AF 1 (20pg/mL; in PBS). and 50pLof AFB,-POD conjugates (20pg/mL in PBS·Tween) was added and mixed. Afterincubation at room temperature for 30 min, the wells were washed and reactedwith 100 pL of the substrate at 37°C. The enzyme reactions were stopped by addi­tion of 50 p of 2 N-HzS04' and the absorbance was measured at 492 nm. The CV(%) of inter-well (n == 4) at 5, 50, and 500 pg/assay were 3.30,2.22, and 5.66 in10% methanol, 5.7, 3.04, and l09 in 30 % methanol, and 2.78, 10.4, and 20.6 in60 % methanol, respectively. The CV (%) of inter-assay (n == 4) at 5, 50, and50 pg/assay were 2.81, 13.7, and 3.88 in 10% methanol, 37.7, 2.52, and 2.19 in 30%methanol, and 4.91, 7.66, and 12.4 in 60% methanol, respectively.

Tab 4: Effects of crude extract from butter peanuts on the direct cE LISA

AFB, dissolved inBuffer" Butter peanut extract''% of binding (CV)C 00492 % of binding (eV)

Equivalent toAFB, (ng{g)

o0.20.51.02.05.0

10.020.050.0

0°492

.551d

.541

.501

.452

.384

.255

.158

.095

.043

100.0 ( 5.28)98.2 ( 4.09)90.9 ( 4.59)82.0 ( 5.18)69.7 ( 7.10)46.2 ( 6.45)28.6 ( 6.45117.3 ( 7.9718.5 (11.90l

.458

.450

.432.374.311.219.135.075.031

100.0 (3.10)98.3 (1.62)94.4 (3.55)81.7 (1.29)68.0 (6.00)47.8 (4.43)29.4 (6.02)16.3 (4.34)6.7 16.36)

• Methanol-water-PBS/Tween = 3:2:5b Non - contaminated butter peanut (10gl was shaken with 50 mL of methanoi­

water (3 : 2) solvent. The extract was spiked with different concentrations ofAFB,. The AFB,·spiked extracts were diluted with an equal volume of PBS /Tween.

C Coefficient variations ICV) were calculated as CV (%) = SO / mean x 100.d Means of six plates. The means of CV (%) of inter· well and inter - assay were

6.27 and 15.3. respec tivelv,

84

Mycotoxin Research, Vol, 4 ('988)

tration of 1.25mg / well, the detection limit is higher than 2 ng/ g. In addition, theextraction efficiency for aflatoxin was low when methanol concentration was lowerthan 60% of the solvent/peanut ratio was lower than 5:1 (26,27).

The next, we examined whether standard curve for AFB] was constructed bythe extracts spiked with AFB] equivalent to 0.2 to 50ng/g and 2-fold diluted withPBS / Tween (final 30 % methanol). As shown in Fig 5 and Table 4, the standardcurve of AFB1 in the buffer (methanol-water-PBS/Tween =30 :20: 50) was actuallythe same as that in the extract solution when the absorbance of control wells addedthe AFB1-free crude extract solution was set at 100% binding. Their correlationcoefficient (= r, n = 6) was calculated to 0.999. The means of the coefficients ofvariations of inter-wells and inter-assays at 0 to 50.0ng of AFB] per g were 3.35(n = 4) and 4.07 (n = 6), respectively. This finding suggested that AFB1 in "butterpeanuts" can be estimated from the standard curve obtained with the buffer solution(Fig 4).

100

­o;-..- 50OJC.-"CC.-m

o

o 0.2 0.5 1 2 5AFB1 (ng/g)

Fig 5: Effect of the crude extract of "butter peanuts" on the standard curve indirect cE LISA.Extracts from AFB,-non-contaminated "butter peanuts" were spiked with AFB,as equivalent to 0.2 - 50 ng/g, and then diluted 2-fold with PBS I Tween. Thesamples (0.05mL) were added to the wells which were coated with 0.1 mL of anti­AFB, MAb AF 1 (200 ng/mL in PBS I Tween) and then 0.05 mL of AFB,-PODconjugate (20,ug/mU were added to the wells. After incubated at room temper­ature for 30 min, the wells were washed and bound enzyme were measured. Thestandard curves of AFB, in the extracts 2-fold diluted with PBS I Tween (e) andthe buffer (0) (methanol - water - PBS I Tween = 3: 2 : 5) were compared at eachconcentration of AFB,. Their correlation coefficient (= r) was 0.999.

85

*

Tab 5: Determination of AFB, in "butter peanuts"

AFB,* Mean ± SO cv** Recovery(ng/g) [n = 3) (%) (%)

1.0 1.2 ± 0.3 22.5 1155.0 4.3 ± 0.6 14.5 86

10.0 9.2± 1.1 12.0 9220.0 21.4±2.0 9.7 10750.0 51.7 ± 4.0 8.7 103

Average 1.6 13.6 101

"Butter peanuts" naturally contaminated with 200 ng/g AFB, were mixed insteps with non-contaminated ones. AFB, was determinated by the directcELISA.

** Coefficient variation (CV) was calculated as follows: CV (%) = SD / mean x 100.

Recovery of AFB, from "butter peanuts" estimated by direct cELISA

The contents of AFB! in "butter peanuts", prepared by mixing the naturally con­taminated sample with non-contaminated one, were estimated by the above directcELISA. The original content of AFB! in the contaminated sample was previouslyanalyzed by HPLC. As summarized in Table 5, AFB! in a range of 5.0 to 50.0ng/gwas estimated with coefficient variation (CV) of about 10%.

It is well known that analytical methods for AFB! using TLC and HPLC requirecomplicated clean-up procedures before the estimation, and are thus time-con­suming. The present direct cELISA using the anti-Al-B, MAb AF I is applicable toestimate 1ng level of AFB! in "butter peanuts" after one-step extraction. The assaytime of this ELISA from weighing the samples to the calculation of contents wasless than 2 hrs. 30 samples could be analyzed within 1hr. Furthermore, the detectionlimit is about 7-fold higher than that proposed by Ram and Hart (15).

The easy extraction procedure, high sensitivity and simple assay procedure couldbe applied to the analysis of other foodstuff for AFB!.

Acknowledgments

This work was supported in part by the Cancer Research Grant from the Ministryof Health and Welfare, Japan, the Japan Private School Promotion Foundation,Tokyo, and the Foundation for Food Science and Technology, Ichikawa. We thankDr Jeo Chiba, National Institute of Health, Tokyo, and Dr Toshitsugu Tanaka,the Public Health Institute of Kobe City, Kobe, for their generous supply of mye­loma cell lines and the samples of "butter peanuts", respectively. The determinationof isotypes of MAbswas kindly carried out by Dr Taizo Uda, Ube Kosan Co, Ube.

References

Heathcote JG (1984) Aflatoxins and related toxins, pp 89 -132. In: V Betina(ed), Mycotoxins, Elsevier,Amsterdam

2 Plamgren MS and Hayes AW (1987) Aflatoxins in food, pp 65-95. In: P Krogh(ed), Mycotoxins in Food, Academic Press, London

86

Mycotoxin Research, Vol. 4 (1988)

3 van Rensburg SJ, Cook-Mozaffari P, van Schalkwyk DJ, van der Watt JJ, Vin­cent TJ, and Purchase IF (1985) Hepatocellular carcinoma and dietary aflatoxinin Mozambique and Transkei. Br J Cancer 51: 713-726

4 Sun T, Wu Y, and Wu S (1983) Monoclonal antibody against aflatoxin B\ andits applications. Clin J Oncol 5: 401- 405

5 Bulatao-Jayme J, Almero EM, Castro MCA, Jardeleza MT, and Salamat LA(1982) A case-control dietary study of primary liver cancer risk from aflatoxinexposure. IntJ Epidemiolll (2):112-119

6 Stoloff L (1983) Aflatoxin as a causeof primary liver-cel1 cancer in the UnitedStates: a probability study. Nutr & Cancer 5 :165-186

7 World Health Organization (1985) Prevention of Liver Cancer, WHO TechnicalReport Ser 691, Geneva

8 Ueno Y (1987) Mycotoxins, pp 139 204. In: K Miller (ed) ToxicologicalAspects of Food, Elsevier, Amsterdam

9 McMahon G, Hanson L, Lee JJ, and Wogan GN (1986) Identification of an acti­vated c- Ki-ras oncogene in rat liver tumors induced by aflatoxin B\. Proc Nat!Acad Sci USA 83 :9418-9422

10 McMahon G, Davis E, and Wogan GN (1987) Characterization of c-Ki-ras onco­genes alleles by direct sequencing of enzymatically amplified DNA from carcino­gen-induced tumors. Proc Natl Acad Sci USA 84 :4974-4978

II Tashiro F, Morimura S, Hayashi K, Makino R, Kawamura H, Horikoshi N,Nemoto K, Ohtsubo K, Sugimura T, and Ueno Y (1986) Expression of thec-Ha-ras and c-myc genes in aflatoxin Bj-induced hepatocellular carcinoma.Biochem Biophys Res Comm 138 :858- 864

12 Chu FS and Ueno I (1977) Production of antibodies against aflatoxin B\. ApplEnviron Microbiol 33 :1125-1128

13 Chu FS, Hsia MTS, and Sun PS (1977) Production of antibodies against afla­toxin B. for enzyme-linked immunosorbent assay. J Assoc Off Anal Chern60:791-794 -

14 El-Nakib, Pestka JJ, and Chu FS (1981) Determination of aflatoxin B\ incom, wheat, and peanut butter by enzyme-linked immunosorbent assay andsolid phase radioimmunoassay. J Assoc Off Anal Chern 64: 1077-1082

15 Ram BP and Hart LP (1986) Enzyme-linked immunosorbent assay of aflatoxinB\ in naturally contaminated corn and cottonseed. J Assoc Off Anal Chern 69:904-907

16 Wild CP, Umbenhauer D, Chpot B, and Montesano R (1986) Monitoring ofindividual human exposure to aflatoxins (AF) and N-nitrosamines (NNO)by immunoassay. J Cell Biochem 30: 171-179

17 Pestka JJ, Li YK, and Chu FS (1982) Reactivity of aflatoxin B2a antibodywith aflatoxin Bj-modified DNA and related metabolites. Appl Environ Micro­bioI44:1159-1165

18 Groopman JD, Trubel U, Donahue PR, Marshak-Rothstien A, and Wogan GN(1985) High affinity monoclonal antibodies for aflatoxins and their applicationto solid-phase immunoassays. Proc Nat! Acad Sci USA 81: 7728-7731

19 Hertzog PJ, Shaw A, Smith L J r, and Garner RC (1983) Improved conditionsfor the production of monoclonal antibodies to carcinogen-modified DNA,for use in enzyme-linked immunosorbcnt assays (ELISA). J Immunol Meth 62:49-58

20 Kohler G and Milstein C (1975) Continuous cultures of fused cells producingantibody of predefined specificity. Nature (London) 256: 495 - 497

21 Oi VI and Herzenburg LA (1980) Immunoglobulin- producing hybrid celllines, pp351-371. In: DB Mishell and SW Shiigi (eds) Selected Methods inCel1ular Immunology. WH Freeman & Co, San Francisco

87

22 Pestka JJ, Gaur PK, and Chu FS (1980) Quantitation of aflatoxin B1 and afla­toxin B1 antibody by an enzyme-linked immunosorbent microassay. ApplEnviron Microbio140: 1027-1031

23 Ishibashi E (1982) The methods of labeling enzyme to the low molecular sub­stances in enzyme immunoassay, pp 128-138. In: E Ishibashi, T Kawai, andK Miyai(eds) Enzyme Immunoassay, Igakushoin, Tokyo (in Japanese)

24 Candlish AAG, Stimson WH, and Smith JE (1985) A monoclonal antibody toaflatoxin B1 : detection of the mycotoxin by enzyme immunoassay. Lett ApplMicrobiol 1: 57- 61

25 Dragsted LO, Bull I, and Autrup H (1988) Substance with affinity to a mono­clonal aflatoxin B1 antibody in Danish urine samples. Fd Chern Toxicol 26:232-242

26 Whitaker TB, Dickens JW, and Giesbrecht FG (1984) Effect of methanol con­centration and solvent: peanut ratio on extraction of aflatoxin from raw peanut.J Assoc Off Anal Chern 67: 35 -36

27 Whitaker TB and Dickens JW (1986) Optimal methanol concentration andsolvent / peanut ratio for extraction of aflatoxin from raw peanuts by modifiedAOAC method II. J Assoc Off Anal Chern 69 :508 - 810

Manuscript received May 19, 1988; accepted July 13, 1988.

88