Vol. 57, No. 3APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1991, p. 836-8420099-2240/91/030836-07$02.00/0Copyright © 1991, American Society for Microbiology

Rapid and Sensitive Sandwich Enzyme-Linked ImmunosorbentAssay for Detection of Staphylococcal Enterotoxin B in Cheese

CELINE MORISSETTE,1 2t* JACQUES GOULET,2 AND GILLES LAMOUREUX'

Centre de Recherche en Immunologie, Institut Armand-Frappier, Laval, Quebec H7V 1B7,1 and Departement de Scienceset Technologie des Aliments, Pavillon Paul-Comtois, Universite Laval, Sainte-Foy, Quebec GIK 7P4, Canada

Received 25 June 1990/Accepted 10 December 1990

A rapid and sensitive screening sandwich enzyme-linked immunosorbent assay (ELISA) was developed forthe detection of staphylococcal enterotoxin B (SEB) in cheese by using a highly avid anti-SEB antibody (Ab) asthe capture Ab (CAb) and as the biotinylated Ab conjugate. The glutaraldehyde fixation method for theimmobilization of CAb on polystyrene dipsticks was superior to the adsorption fixation and the adsorption-glutaraldehyde fixation methods. The glutaraldehyde fixation method resulted in a higher surface-saturatingCAb concentration as evaluated by the peroxidase saturation technique and by the ability of the CAb-coateddipstick to discriminate between positive and negative controls (index of discrimination). Of nine blockingagents used alone or in pairs, lysine-human serum albumin, bovine serum albumin, human serum albumin,and gelatin effectively saturated available sites on the CAb-coated dipsticks without causing interference withthe antigen-Ab reactions. The addition of 1% polyethylene glycol to the diluent of the biotinylated anti-SEB Abconjugate improved the detection of SEB. A concentration of 4% polyethylene glycol allowed a 5-min reactiontime for the streptavidin-biotin-horseradish peroxidase conjugate. Cheddar cheese homogenate reduced thesensitivity of the SEB assay; however, the sensitivity was restored when 1.6% (wt/vol) of either a nonionicdetergent (Mega-9) or two zwitterionic detergents (Zwittergent 3-10 and 3-12 detergent) was added to thediluent. By using the rapid sandwich ELISA, a minimum of 0.5 to 1.0 ng of SEB per ml was detected within45 min. The whole procedure for the analysis of the cheddar cheese samples was completed within 1 h. Thissandwich ELISA could be a rapid and sensitive screening method for the detection of staphylococcalenterotoxins in food for the agri-food industries.

Staphylococcal enterotoxins (SE) are one of the leadingcauses of bacterial food poisoning in Canada (36). Thesetoxins in the active state are resistant to both proteolyticenzymes, such as trypsin, chymotrypsin, rennin, and papain(2), and high temperatures (boiling crude solutions for 30min) (3). They have been found in organic raw materials aswell as pasteurized, cooked, or otherwise processed foods(25). Rapid screening of SE for quality control in the foodindustry relies on rapid and sensitive methods.Many enzyme immunoassay techniques have been de-

scribed for the detection of SE in culture supernatants and infood samples (1, 9-11, 13, 18, 24, 26, 30, 32, 37-39). Theenzyme-linked immunosorbent assay (ELISA) is as efficientand sensitive as the radioimmunoassay but does not requirethe use of radioactive materials. Four types of ELISA werestudied for their ability to detect SE in food samples (9).Whereas the competitive ELISA showed a higher specificityand was not sensitive to protein A in food extracts, thesandwich ELISA with labeled antibody (Ab) gave the bestoverall performance.Sandwich ELISAs are commonly performed by immobi-

lizing a capture Ab (CAb) on plastic supports. The antigen(Ag) captured by the CAb support may be detected either byan enzyme-labeled Ab specific for the same determinant asthe CAb or by an enzyme-labeled Ab recognizing a differentepitope on the captured, multivalent Ag (4, 5). Effectivepairs of monoclonal Abs and polyclonal Abs can be used inthe sandwich ELISA to detect SE (20, 34). The performance

* Corresponding author.t Present address: Food Research Centre, Research Branch,

Agriculture Canada, Central Experimental Farm, Ottawa, OntarioKlA 0C6, Canada.

of the ELISA depends on the apparent affinity (avidity) andspecificity of the selected Ab.

This article describes the development of a sandwichELISA to be used as a rapid screening assay for thedetection of SE in food. By using avid polyclonal Abs tostaphylococcal enterotoxin B (SEB) and a biotin-streptavi-din amplification system, each major step of the analyticalprocedure (immobilization method of CAb, diluent for re-agents, incubation period, and diluent for food preparation)was optimized and standardized to improve the sensitivityand the reaction rate of the test.

MATERIALS AND METHODS

SEB and Ab to SEB. SEB (Sigma Chemical Co., St. Louis,Mo.; lots 32F-4024 and 94F-4001) was used as the Ag. Thepreparation was rehydrated in sterile distilled water at aconcentration of 1.5 mg of SEB per ml. Rabbit anti-SEB Abcontaining the crude immunoglobulin fraction of antiserumwas purchased from Sigma (lot 91F-40171).

Diluents. Phosphate-buffered saline (0.01 M sodium phos-phate-0.15 M NaCl, pH 7.4) (PBS) containing 0.02% sodiumazide (NaN3) and supplemented with 0.05% Tween 20, 0.1%bovine serum albumin (BSA), and 4% polyethylene glycol6000 (PEG) was used as a general diluent. In some experi-ments, the concentration of PEG was varied. Physiologicalsaline containing 0.9% NaCl, 0.05% Tween 20, and 0.01%NaN3 was used as a washing solution.

Biotinylated Abs. The anti-SEB Ab was biotinylated byuse of the procedure of GIBCO-BRL (Burlington, Ontario,Canada). Biotin-N-hydroxysuccinimide ester (Bethesda Re-search Laboratories/Life Technologies, Inc., Gaithersburg,Md.; lot 51101) was dissolved to a concentration of 50 mg/ml

836

on August 3, 2020 by guest

http://aem.asm

.org/D

ownloaded from

RAPID DETECTION OF SEB IN CHEESE 837

TABLE 1. Molecular weight, critical micellar concentration, andionic state of detergents testeda

Detergent Mol Critical micellar Ionic statewt concn (M)

Detergent test kitHexyl-p-D-glucopyranoside 264.3 Large NonionicHeptyl-3-D-glucopyranoside 278.3 NonionicOctyl-f3-D-glucopyranoside 292.4 2.5 x 10-2 NonionicNonyl-p-D-glucopyranoside 306.4 6.5 x 10-3 NonionicDodecyl-,B-D-maltoside 510.6 Nonionic

Mega test kitMega-8 321.5 NonionicMega-9 335.5 Nonionic

Zwittergent test kitZwittergent 3-08 detergent 279.6 Large ZwitterionicZwittergent 3-10 detergent 307.6 3.9 x 10-2 ZwitterionicZwittergent 3-12 detergent 335.6 3.6 x 10-3 ZwitterionicZwittergent 3-14 detergent 363.6 3.3 x 10-4 Zwitterionic

Other common detergentsSodium dodecyl sulfate 288.4 IonicTween 20 5.9 x 10-5 NonionicDishwashing liquid (Ivory)

a From Calbiochem Biochemical & Immunochemical catalog 87, Heleniuset al. (14), and Gardas and Lewartowska (12).

in dry N,N-dimethyl formamide (Fisher Scientific Co., Or-angeburg, N.Y.; catalog no. D-119). A 10-pul volume of thissolution was reacted with 0.125 ml of protein (20 mg/ml), 20p.l of 0.5 M NaHCO3 (pH 9.0), and 40 p.l of H20 for 1 h atroom temperature with gentle agitation. The reaction wasstopped by adding 25 p. of 1 M NH4Cl. The mixture wasdialyzed against PBS and membrane filtered (pore size, 0.45Am).A biotinylated donkey anti-rabbit immunoglobulin (Amer-

sham Canada Ltd., Oakville, Ontario, Canada; catalog no.RPN 1004, lot 30) was also used to detect the presence ofrabbit Abs on dipsticks.

Streptavidin-biotinylated enzyme complex. Biotinylated an-ti-rabbit immunoglobulin and biotinylated anti-SEB Ab weredetected by a preformed streptavidin-biotinylated horserad-ish peroxidase complex from Amersham Canada (catalog no.RPN 1051, lot 42 and 61).Enzyme substrate. A horseradish peroxidase substrate

solution was freshly prepared by dissolving 100 mg ofo-diaminobenzene dihydrochloride (Sigma) in 100 ml of 0.2M Na2-phosphate-0.1 M citrate buffer, pH 5, containing 150p.l of 30% H202. The peroxidase reaction was stopped by theaddition of 0.5 ml of a solution containing 2 M H2SO4 and 0.1M NaSO3 (27).

Detergents. Several detergents (Table 1) were used assubstitutes for Tween 20 in the diluent for the preparation offood homogenates. All selected detergents were soluble inPBS-NaN3-BSA-4% PEG. The Detergent Test Kit, MegaTest Kit, and Zwittergent Test Kit were obtained fromCalbiochem Biochemical & Immunochemical (TerochemLaboratories Ltd., Edmonton, Alberta, Canada). Sodiumdodecyl sulfate was obtained from Matheson Coleman &Bell, Norwood, Ohio, and Tween 20, polyoxyethylene sor-bitan monolaureate, was obtained from Fisher Scientific.

Immobilization of CAbs on polystyrene dipsticks. (i) Ad-sorption fixation method. A batch of 500 polystyrene dip-sticks (Sarstedt no. 81,970, Ruhr, West Germany) waswashed with deionized water for 2 h and incubated overnight

in a 0.01 M carbonate buffer (pH 9.6) containing increasingconcentrations (0, 0.5, 1, 5, 10, and 25 p.g/ml) of CAb. Insome experiments, the remaining active sites on the CAb-coated dipsticks were saturated for 2 h with a solution ofblocking agent (1% BSA in PBS). All incubations wereperformed at room temperature (22°C) with agitation on arotary shaker. After two washings with PBS-NaN3-Tween20, the CAb-coated dipsticks were stored in PBS-NaN3-Tween 20 at 4°C until required.

(ii) Combined adsorption-glutaraldehyde fixation method.The adsorption-glutaraldehyde fixation method was adaptedfrom the binding technique of Saunders and Clinard (31). Abatch of 500 polystyrene dipsticks previously washed withdeionized water for 2 h was incubated for 1 h in a 0.01 Mcarbonate buffer (pH 9.6) containing increasing concentra-tions (0, 0.5, 1, 5, 10, and 25 pug/ml) of CAb. The CAb-coateddipsticks were then incubated with 0.25% (vol/vol) glutaral-dehyde in PBS (pH 7.0) for 30 min, washed four times withdistilled water, and reincubated overnight in the same CAb-containing solutions. In some experiments, the remainingactive sites of the CAb-coated dipsticks were saturated witha solution of blocking agent (1% BSA in PBS) for 2 h. Allincubations were performed at room temperature (22°C) withagitation on a rotary shaker. After two washings withPBS-NaN3-Tween 20, the CAb-coated dipsticks were storedin PBS-NaN3-Tween 20 at 4°C until required.

(iii) Glutaraldehyde fixation method. The glutaraldehydefixation method was adapted from the method of Suter (33)by the method of Tijssen (35). A batch of 500 polystyrenedipsticks was washed twice with distilled water and oncewith 0.1 M sodium phosphate buffer (pH 5). The dipstickswere pretreated with 0.2% glutaraldehyde (Sigma; catalogno. G-6257, lot 116F-5027, grade II) in 0.1 M sodiumphosphate buffer (pH 5) for 4 h. The dipsticks were washedthree times with the same buffer and incubated in 0.1 Msodium phosphate buffer (pH 8) containing increasing con-centrations (0, 2, 4, 8, 10, 20, 40, and 60 p.g/ml) of CAb for 14to 16 h. The CAb-coated dipsticks were then washed threetimes with PBS (pH 7.4), and in some experiments thenon-Ab-occupied sites were saturated by incubating dip-sticks for 2 h in a solution of blocking agent (100 mM lysine)in PBS. All incubations were performed at room temperature(22°C) with agitation on a rotary shaker. After two washingswith PBS (pH 7.4), the dipsticks were stored in PBS-NaN3-Tween 20 at 4°C until required.Rapid sandwich ELISA. In a typical assay (Fig. 1), a

CAb-coated dipstick was incubated with 1.0 ml of SE-contaminated sample for 5 min. All incubation steps wereperformed at room temperature with agitation. After twowashings with the washing solution, the CAb-coated dipstickwas transferred to a tube containing 300 p.1 of a dilution ofbiotinylated anti-SE Ab conjugates. After 5 min of incuba-tion, the dipstick was washed twice and transferred to a tubecontaining 300 pul of diluted streptavidin-biotin-horseradishperoxidase complex. The CAb-coated dipstick was removedafter 5 min, washed twice, and transferred to a tube contain-ing 1.0 ml of H202-o-diaminobenzene dihydrochloride for 5min. After removing the dipstick, the reaction was stoppedby transferring the reaction mixture into tubes containing thestopping solution. The optical density of the solution at 490nm (OD49) was read with an optical fiber probe (PC 800;Brinkmann [Canada Ltd.], Rexdale, Ontario, Canada).

Blocking treatment. Several reagents (Table 2) were testedfor their ability to block nonspecific binding and saturate theunoccupied sites on surface-immobilized CAb. They were

used alone or in pairs at a concentration of 1% in PBS. After

VOL . 57 i 1991

on August 3, 2020 by guest

http://aem.asm

.org/D

ownloaded from

838 MORISSETTE ET AL.

Step 1

s...,..0 0 Io

0

C'0. ,S

20 minutes Washing

Step 2

10 minutes Washing

Step 3

-4

Washing

OD 490

* SEA Antibody against SEA Biotinylated anti-SE Ab conjugate

A _ s_%b _A _ _ t ._s2_

FIG. 1. Major steps of the rapid sandwich ELISA procedure.OPD, o-Diaminobenzene dihydrochloride.

immobilization by using an unsaturating concentration ofCAb (4 ,ug/ml), the dipsticks were incubated with the block-ing agents for 2 h and washed three times. The capacity ofthe blocking agents to saturate the Ab-free sites on CAb-coated dipsticks was evaluated by the peroxidase saturationtechnique. The rapid sandwich ELISA with the blocked

TABLE 2. Reagents used as blocking agents after immobilizationof Abs on polystyrene dipsticks

Reagent Source

DL-Lysine ................ Sigma (catalog no. L-6001)Glycine ................. Fisher (catalog no. G-46)Human serum albumin .........In-house preparation (Institut

Armand-Frappier)BSA ................. Miles Laboratories, Inc., Naperville,

Ill. (catalog no. 81-066-1)Whole rabbit serum.............In-house preparation (single bleed

from one animal)Bacto-Peptone ................ Difco (catalog no. 0118-05-4)Casein ................ Sigma (catalog no. C-5890)Gelatin ................ Sigma; from calf skin type IV

(catalog no. G-0510)Organosilane (Prosil-28) .......SCM Chemicals, Gainesville, Fla.

(catalog no. 11975-0)

CAb-coated dipsticks was employed to discriminate be-tween positive and negative controls.

Peroxidase saturation technique. The peroxidase saturationtechnique of Munioz et al. (22) was used to estimate theAb-free surface sites after the immobilization. Horseradishperoxidase (Boehringer GmbH; lot 10300224-40, 308 U/mg)diluted to 0.1 mg/ml in PBS was incubated with CAb-coateddipsticks for 2 h at room temperature with agitation. Thedipsticks were then washed four times with PBS, and theenzyme activity was developed as described above.

Preparation of cheese samples. Five-gram samples of ched-dar cheese contaminated with 200 ng of SEB per g werehomogenized in a blender for 15 s with 10 ml of diluentcontaining various concentrations (0.016, 0.16, and 1.6%[wt/vol]) of detergents (Table 1). A 1.0-ml aliquot of thehomogenate was used for testing with the rapid sandwichELISA.

Index of discrimination. To evaluate the immunoreactivityof the CAb-coated dipsticks in the rapid sandwich ELISA,we used a discrimination index defined as follows: (OD490 ofpositive control - OD490 of negative control)/OD490 ofnegative control x 100 = %.

RESULTSSelection of Abs. The anti-SEB Ab used in this study was

selected for its high avidity for SEB. The anti-SEB AbELISA titer of the crude immunoglobulin fraction was409,600 at the endpoint of negative control, when incubatedfor a 5-min period with SEB-coated dipsticks and a dilutionof the protein A-horseradish peroxidase conjugate (21).

Selection of a method to immobilize the CAb. Three immo-bilization methods, adsorption fixation, adsorption-glutaral-dehyde fixation, and glutaraldehyde fixation, were comparedfor their ability to bind CAb to polystyrene dipsticks undersaturating conditions and for the efficiency of the resultingCAb-coated dipsticks to discriminate between negative andpositive controls. After the immobilization of CAb, residualfree sites were evaluated by the peroxidase saturation tech-nique. In this technique, the OD value is directly propor-tional to the available binding sites on support. Accordingly,the OD values decreased as the CAb concentration in-creased, and regression lines for the rapidly decreasing ODreadings and for the slowly decreasing (near-constant) ODreadings were calculated (Fig. 2). By using the glutaralde-hyde fixation method, the surface-saturating CAb concentra-tion was evaluated as the intersection point between bothregression lines (22), i.e., 11 ,ug/ml. The adsorption fixation

_

tt':..;

.-.

A:..:.-.

..

o

APPL. ENVIRON. MICROBIOL.

on August 3, 2020 by guest

http://aem.asm

.org/D

ownloaded from

RAPID DETECTION OF SEB IN CHEESE 839

0.7E

0.6

z 0.5z

< 0.40.400.300.3 .. .. .. ..

0 10 20 30 40 50 60CONCENTRATION OF CAb (ligtmQ

FIG. 2. Immobilization of increasing concentrations of CAb withthe glutaraldehyde fixation method. Non-Ab-occupied surface siteswere evaluated by the peroxidase saturation technique (n = 8,arithmetic mean of OD values from a test repeated four times induplicate).

and adsorption-glutaraldehyde fixation methods demon-strated a surface-saturating CAb concentration of approxi-mately 5 ,ug/ml, a twofold decrease compared with theconcentration demonstrated by the glutaraldehyde fixationmethod. The ability to discriminate between positive andnegative controls with the CAb-coated dipsticks obtainedfrom the glutaraldehyde fixation method is represented inFig. 3 and 4. A maximum immunoreactivity was observed atabout 10 jig of CAb per ml. Results expressed as the index ofdiscrimination allow visualization of the surface-saturatingCAb concentration where the OD value of the positivecontrol is increased while the OD value of the negativecontrol remains low. The index of discrimination for theadsorption fixation and the adsorption-glutaraldehyde fixa-tion methods was at best 94 and 265%, respectively, at aconcentration of 5 ,ug of CAb per ml. The adsorption-glutaraldehyde fixation and the glutaraldehyde fixation meth-ods gave good discrimination at their surface-saturating CAbconcentration; however, the glutaraldehyde fixation methodwas preferred because of its higher index of discriminationand its higher surface-saturating CAb concentration. How-

1.2 Negative controlE ---& GA-CAb

010 GA-CAb-Lysne* Positive control

F _iir ~GACAb_< 0.8 - GA-CAb-Lysin

0.6w

..j 0.4

0

0 10 20 30 40 50 60CONCENTRATION OF CAb (1tgIml)

FIG. 3. Immunoreactivity of CAb-coated dipsticks evaluated bythe detection of a positive control (laboratory diluent containing 145ng of SEB per ml) and a negative control (laboratory diluent) byrapid sandwich ELISA (arithmetic mean of OD values from a testperformed in duplicate). GA, Glutaraldehyde fixation method.

900,2 800

O 700

> 6002 500a:) 400

D 3000 2000 100z

0

GA-CAb

v GA-CAb-Lysirw

0 10 20 30 40 50 60CONCENTRATION OF CAb (igml)

FIG. 4. Immunoreactivity of CAb-coated dipsticks as estimatedby the index of discrimination. GA, Glutaraldehyde fixation method.

ever, when this latter method was used with the suggestedblocking agent, lysine (35), interference was observed on thedetection of SEB (Fig. 3 and 4).

Selection of a blocking agent. Preliminary tests were con-ducted with a series of nine blocking agents used alone or inpairs and with dipsticks coated with 4 jig of anti-SEB Ab perml (undersaturating CAb concentration). Table 3 shows thecapacity of the four best blocking agents to inhibit thebinding of the peroxidase and their influence over theimmunoreactivity of the CAb-coated dipsticks in the rapidsandwich ELISA. The index of discrimination of the CAb-coated dipsticks blocked with those agents was similar to theone obtained with the unblocked controls, indicating thattheir effect is not detrimental to the immunological reactionin undersaturating conditions. The following other blockingagents gave either a high OD reading with the peroxidasesaturation technique (do not saturate available sites) or poorindex of discrimination (interfere with the further reactions):lysine, lysine-organosilane, lysine-BSA, lysine-glycine, gly-cine, glycine-organosilane, glycine-human serum albumin,peptone, peptone-lysine, peptone-glycine, and whole rabbitserum. At the optimal coating CAb concentration (11 ,ug ofanti-SEB Ab per ml), the effect of the four best blockingagents was maintained. Because of its high index of discrim-ination and its low OD value for negative control, gelatin wasselected as the best blocking agent for high dilutions ofbiotinylated anti-SEB Ab.

Selection of the incubation periods. Minimal incubationperiods were selected to detect the critical enterotoxinconcentration (100 ng of SE per 100 g of food or 0.5 ng of SEper ml of food homogenate) assuming a dilution factor of 2.This minimum concentration was detected when the periodof incubation of CAb-coated dipstick with the SEB-contain-ing sample was longer than 10 min; for an incubation time of20 min, the minimal detection limit is around 0.5 ng of SEBper ml (Table 4). This period of incubation can be increasedwhen the dilution factor is higher than 2.

Effect of PEG on immune reaction. The addition of 0.5 to4% PEG to the diluent of biotinylated anti-SEB Ab resultedin elevated background OD readings for concentrationsabove 1% (Fig. 5). Because PEG improved the detection ofthe positive control, a concentration of 1% was selected forthe supplementation of the biotinylated anti-SEB Ab diluent.

VOL. 57, 1991

on August 3, 2020 by guest

http://aem.asm

.org/D

ownloaded from

840 MORISSETTE ET AL.

TABLE 3. Saturation level and immunoreactivity of blocked CAb-coated dipsticks

Peroxidase binding Detection of SEB (OD value) (n = 4)Dipsticks (OD value) Pos Cb Index of

(n6) Neg C' ~~~~~~~~~~~~~~discrimination (%Controls

Untreated 1.309 ± 0.078 0.001 0.001GA treated 0.898 ± 0.120 0.004 0.006

CAb-coated dipsticks (GA)CUnblocked 1.018 ± 0.104 0.125 ± 0.019 0.436 ± 0.112 249Blocked with 1% lysine-1% HSAd 0.318 ± 0.090 0.119 ± 0.012 0.470 ± 0.079 295Blocked with 1% BSA 0.138 + 0.052 0.165 ± 0.023 0.557 ± 0.029 238Blocked with 1% HSAd 0.291 ± 0.057 0.141 ± 0.007 0.473 ± 0.031 235Blocked with 1% gelatin 0.325 ± 0.048 0.103 ± 0.009 0.409 ± 0.030 297a Negative control: laboratory diluent without SEB.b Positive control: laboratory diluent with 100 ng of SEB per ml.c GA, Glutaraldehyde fixation method.d HSA, Human serum albumin.

Varying the PEG concentration in the sample diluent didnot affect the discrimination between positive and negativecontrols, while addition of 4% PEG was required to obtain areaction with the streptavidin-biotin-horseradish peroxidasecomplex in a 5-min incubation period (results not shown).

Selection of detergents for food homogenates. Of 14 deter-gents tested for their ability to solubilize milk fat, only 3were able to restore the OD readings necessary for thedetection of SEB in cheddar cheese homogenate (higher than0.4) (Table 5). The other detergents gave OD values equal toor greater (up to 0.3) than that of the control cheese-nodetergent sample. Because of the viscosity of the cheesehomogenate, a dilution factor of 3 was required for thehomogenization, bringing the probable final concentration ofSEB in the liquid phase of the homogenate to approximately66 ng/ml. The OD readings obtained with a 1.6% (wt/vol)solution of these three best detergents correspond approxi-mately to two-thirds of the OD readings obtained with thepositive controls (no cheese-no detergent) and containing100 ng of SEB per ml. Mega-9 and Zwittergent 3-10 and 3-12detergents were successful in eliminating the nonspecificreactions with the CAb-coated dipstick.

DISCUSSION

The reactivity of Abs against immobilized and/or solubleAgs is of primary importance in the development of rapidimmunochemical tests for the food industry. In this work, arapid sandwich ELISA was developed on the basis of thehigh reactivity of an anti-SEB Ab. This Ab gave higherreactivity toward the immobilized Ag than toward the solu-ble Ag (results not shown). The selection of this avidanti-SEB Ab and the optimization of the rapid sandwich

ELISA procedure allowed the detection of 0.5 ng of SEB perml within 45 min.

Immobilization of CAb. Most proteins adsorb to plasticsurfaces very likely through hydrophobic interactions ratherthan covalent binding. Adsorbed proteins may undergodenaturation and lose immunological activity (8). Adsorptionof proteins to polystyrene is concentration dependent up toa saturation limit and can result in substantial loss ofantigenic or Ab activity. Desorption is also observed at asteady rate and can negatively influence the performance ofthe immunoassay (4).

Since desorption rates as high as 40% were observed withAb adsorbed to plastic in the presence of serum samples (7),food samples are likely to show major variations due to theirhighly complex nature (protein and fat composition andconcentration and various other compounds).To standardize and optimize the immunoreactivity of the

immobilized Abs, a conventional adsorption fixation methodand two glutaraldehyde fixation procedures were evaluated.A higher concentration of CAbs was immobilized on poly-styrene dipsticks by glutaraldehyde pretreatment than byconventional adsorption or combined adsorption-glutaralde-hyde fixation procedures. Dobbins Place and Schroeder (7)reported that the glutaraldehyde pretreatment improved Abretention on polystyrene.The Ag capture capacity of the immobilized Ab is critical

for the performance of the assay (5). In this study, theimmobilization of anti-SEB Ab on dipsticks was improvedthrough glutaraldehyde fixation, but no attempts were madeto measure concentrations of immobilized Abs nor to esti-mate adsorption and desorption phenomena.PEG supplementation. The enhancing effect of polymers

on immunological precipitation reactions has been reported

TABLE 4. Effect of incubation time and sequence on the detection level of SEB

Incubation periods OD value of OD value at concn of SEB (ng/ml) of:(min)a negative control 0.5 1.0 10 50 100

10/10/5/5 (n = 6) 0.060 ± 0.011 0.074 Q,Q92b 0.221 0.421 0.54610/10/5/10 (n = 2) 0.078 ± 0.005 0.075 0.115 0.307 0.594 0.72720/10/5/5 (n = 4) 0.068 ± 0.007 Q.089 0.117 0.276 0.538 0.5820/10/5/10 (n = 2) 0.098 ± 0.012 0.121 0.156 0399 0.646 0.718

a Periods correspond to the incubation time with Ag-containing sample/biotinylated Ab conjugate/streptavidin-biotinylated enzyme complex/enzyme substrate.b Underlined values indicate OD readings equal to or higher than the negative control plus 2 standard deviations.

APPL. ENVIRON. MICROBIOL.

on August 3, 2020 by guest

http://aem.asm

.org/D

ownloaded from

RAPID DETECTION OF SEB IN CHEESE 841

0.7 msuv cuo. _ o_ 8pgCAbff JE lzgCAM

0.6 Positive controlQ~~~~~ 0 5g °M40

Z0.3

0.3 ~~~~~~~~~~~~~~~#.100.2

0.1

0 1 2 3 4CONCENTRAMON OF PEG (%)

FIG. 5. Influence of increasing concentrations of PEG in bioti-nylated anti-SEB Ab conjugate diluent over the Ag-Ab reaction.Immunoreactivity of CAb-coated dipsticks, coated with 8- and10-,ug CAb concentrations without blocking treatment, was evalu-ated by discriminating between a positive control (diluent with 100ng of SEB per ml) and a negative control (diluent) by using the rapidsandwich ELISA (arithmetic mean of OD values from a testperformed in duplicate).

(15, 29). In the latter study, inclusion ofPEG (4% [wt/vol]) inthe conjugate diluent enhanced the reaction between theconjugate and the Ag-bound Ab about threefold, both atroom temperature and at 37°C. When PEG was present, a1-h incubation period at room temperature gave OD values

TABLE 5. Effect of detergents on the detection of SEB incheddar cheese samples

OD value at concn ofSample (n = 2) Contamination detergent (wt/vol) of:

1.6% 0.16% 0.016%

No cheese-no detergentDiluentaNeg Cb 0.076Pos Cc 0.748

Laboratory diluentdNeg Cb 0.081Pos C' 0.754

Cheese-no detergentDiluenta _e 0.031

+f 0.178Laboratory diluentd - 0.019

+ 0.141Cheese + Mega-9 - 0.081 0.031 0.004

+ 0.543 0.100 0.168Cheese + Zwittergent - 0.082 0.026 0.041

(3-10 detergent)+ 0.449 0.143 0.171

Cheese + Zwittergent - 0.051 0.028 0.012(3-12 detergent)

+ 0.410 0.134 0.181

a Diluent without Tween 20 consisted of PBS + 0.02% NaN3 + 0.1% BSA+ 4% PEG.

b Negative control consisted of diluent without SEB.Positive control consisted of diluent with 100 ng of SEB per ml.

d Laboratory diluent consisted of PBS + 0.02% NaN3 + 0.1% BSA +0.05% Tween 20 + 4% PEG.eWithout SEB.f Containing 66 ng of SEB per ml of food homogenate.

close to those obtained overnight without PEG (29). Otherpolymers such as dextran and polyvinyl chloride exertsimilar promoting effects on the interaction between solubleAgs and Abs (16). This effect is somewhat related to stericexclusion of the immune complex from the domain of thepolymer (28).The concentration of PEG added to the reagents may vary

according to the Ag-Ab system used. A sandwich techniquewas greatly improved by the addition of 2% PEG to thereagents; a sixfold shortening of the incubation time wasreported by Katnik et al. (17) with no loss in either sensitiv-ity or accuracy.

In our assay, the addition of 4% PEG to the diluent ofbiotinylated anti-SEB Ab increased the background ODvalues, and no evidence could be obtained that concentra-tions higher than 1% could improve the sensitivity of thedetection of SEB.

Effect of detergents. The extraction and concentration ofSE from the food sample is an essential and laborious step inany enzyme immunoassay (9, 11, 19, 23, 37). Since foods arecomplex systems where hydrophobic interactions are com-mon (lipid-lipid, protein-protein, lipid-protein), detergentsare likely to help minimize interference by these constituentsreacting with the immunoreagents. Nonionic detergents canthus be added to prevent hydrophobic interactions betweenadded proteins and the solid phase but without disrupting thefixation of the Ag or Ab previously adsorbed to plasticsurfaces (6).The concentration of detergent needed to overcome hy-

drophobic interactions in aqueous systems depends on thecritical micellar concentration, the micelle size, the temper-ature, the nature of the system, and the detergent (14). Inaqueous solutions, detergent molecules occur in the form ofmonomers and micelles. The micelles are fairly monodis-perse compact aggregates where apolar groups are seques-tered in the center and polar groups face outward.The major extraction problem that was observed in our

test system was the adhesion of fatty material to the CAb-coated dipsticks. After testing a full range of detergents(Table 1), one nonionic (Mega-9), and two zwitterionic(Zwittergent 3-10 and 3-12 detergents) detergents wereshown to restore the Ag-Ab reaction at the level of thepositive control in the absence of cheese and detergents.Further studies need to be undertaken with other foods toestablish optimum detergent concentrations to be used tomaximize the sensitivity of the test.

Sensitivity. The double-antibody sandwich ELISA devel-oped by Freed et al. (11) allowed the detection of less than 1ng of SE per g of sample in several types of foods. Thedetection procedure could be completed within 3.5 h andnecessitated at least 45 min of food preparation. By usingthis sandwich ELISA and two highly effective murine mono-clonal Abs, Thompson et al. (34) found that the monoclonaland the polyclonal Ab system compared favorably in thedetection of 1.0 ng of SE per ml. Lapeyre et al. (20)developed an indirect double sandwich ELISA by using apair of Abs (monospecific monoclonal Ab as the CAb andpolyspecific rabbit Abs as the detection system) to detect 1ng of SE per ml within 3 h.The rapid sandwich ELISA that is described in this work

gives a detection limit close to what is reported by Freed etal. (11) and Lapeyre et al. (20) but within a very short periodof time (45 min). Indeed, the whole procedure for theanalysis of a food sample can be completed within 1 h. Thissandwich ELISA is thus a rapid and effective screeningmethod for the detection ofSEB in cheese samples, and such

VOL. 57, 1991

on August 3, 2020 by guest

http://aem.asm

.org/D

ownloaded from

842 MORISSETTE ET AL.

a test could be used advantageously by food industries forthe detection of SEs in foods.

ACKNOWLEDGMENTS

The present study was supported by a grant from the Conseil desrecherches et services agricoles du Quebec (grant IAF-84-B-1105).

REFERENCES1. Berdal, B. P., 0. Olsvik, and T. Omland. 1981. A sandwich

ELISA method for detection of Staphylococcus aureus entero-toxins. Acta Pathol. Microbiol. Scand. Sect. B 89:411-415.

2. Bergdoll, M. S. 1970. Enterotoxins, p. 265-326. In T. C. Montie,S. Kadis, and S. J. AjI (ed.), Microbial toxins, vol. 3. AcademicPress, Inc., London.

3. Bergdoll, M. S., R. N. Robbins, K. Weiss, C. R. Borja, I.-Y.Huang, and F. S. Chu. 1973. The staphylococcal enterotoxins:similarities. Contrib. Microbiol. Immunol. 1:390-396.

4. Butler, J. E., J. H. Peterman, M. Suter, and S. E. Dierks. 1987.The immunochemistry of solid-phase sandwich enzyme-linkedimmunosorbent assays. Fed. Proc. 46:2548-2556.

5. Butler, J. E., J. E. Spradling, M. Suter, S. E. Dierks, H.Heyermann, and J. H. Peterman. 1986. The immunochemistry ofsandwich ELISAs. I. The binding characteristics of immuno-globulins to monoclonal and polyclonal capture antibodies ad-sorbed on plastic and their detection by symmetrical andasymmetrical antibody-enzyme conjugates. Mol. Immunol. 23:971-982.

6. Clark, B. R., and E. Engvall. 1980. Enzyme-linked immunosor-bent assay (ELISA): theoretical and practical aspects, p. 167-179. In E. T. Maggio (ed.), Enzyme-immunoassay, chapter 8.CRC Press, Boca Raton, Fla.

7. Dobbins Place, J., and H. R. Schroeder. 1982. The fixation ofanti-HBsAg on plastic surfaces. J. Immunol. Methods 48:251-260.

8. Engvall, E. 1980. Enzyme immunoassay ELISA and EMIT.Methods Enzymol. 70:419-439.

9. Fey, H., H. Pfister, and 0. Ruegg. 1984. Comparative evaluationof different enzyme-linked immunosorbent assay systems forthe detection of staphylococcal enterotoxins A, B, C, and D. J.Clin. Microbiol. 19:34-38.

10. Fey, H., and G. Stiffier-Rosenberg. 1977. Detection of staphylo-coccal enterotoxin B with a new modification of the enzymelinked immunoassay (ELISA). Experientia 33:1678.

11. Freed, R. C., M. L. Evenson, R. F. Reiser, and M. S. Bergdoll.1982. Enzyme-linked immunosorbent assay for detection ofstaphylococcal enterotoxins in foods. Appl. Environ. Micro-biol. 44:1349-1355.

12. Gardas, A., and A. Lewartowska. 1988. Coating of proteins topolystyrene ELISA plates in the presence of detergents. J.Immunol. Methods 106:251-255.

13. Hahn, I. F., P. Pickenhahn, W. Lenz, and H. Brandis. 1986. Anavidin-biotin ELISA for the detection of staphylococcal entero-toxins A and B. J. Immunol. Methods 92:25-29.

14. Helenius, A., D. R. McCaslin, E. Fries, and C. Tanford. 1979.Properties of detergents. Methods Enzymol. 56:734-749.

15. Hellsing, K. 1972. Letter to the editor. Immunochemistry 9:753.16. Hellsing, K., and W. Richter. 1974. Immunochemical quantita-

tion of dextran by a polymer enhanced nephelometric proce-dure. J. Immunol. Methods 5:147-151.

17. Katnik, I., M. Podg6rska, and W. Dobryszycka. 1987. Polyeth-ylene glycol enzyme immunoassay for screening anti-haptoglo-bin monoclonal antibodies. J. Immunol. Methods 102:279-282.

18. Kauffman, P. E. 1980. Enzyme immunoassay for staphylococcalenterotoxin A. J. Assoc. Off. Anal. Chem. 63:1138-1143.

19. Koper, J. W., A. M. Hagenaars, and S. Notermans. 1980.Prevention of cross-reactions in the enzyme linked immunosor-bent assay (ELISA) for the detection of Staphylococcus aureusenterotoxin type B in culture filtrates and foods. J. Food Safety2:35-45.

20. Lapeyre, C., S. V. Kaveri, F. Janin, and A. D. Strosberg. 1987.Production and characterization of monoclonal antibodies to

staphylococcal enterotoxins: use in immunodetection and im-munopurification. Mol. Immunol. 24:1243-1254.

21. Morissette, C., J. Goulet, and G. Lamoureux. 1990. Productionof avid rabbit antibodies against staphylococcal enterotoxins Aand B. J. Food Prot. 53:782-785.

22. Munoz, C., A. Nieto, A. Gaya, J. Martinez, and J. Vives. 1986.New experimental criteria for optimization of solid-phase anti-gen concentration and stability in ELISA. J. Immunol. Methods94:137-144.

23. Notermans, S., R. Boot, P. D. Tips, and M. D. De NooiJ. 1983.Extraction of staphylococcal enterotoxins (SE) from mincedmeat and subsequent detection of SE with enzyme-linked im-munosorbent assay (ELISA). J. Food Prot. 46:238-241.

24. Notermans, S., H. L. Verjans, J. Bol, and M. van Schothorst.1978. Enzyme linked immunosorbent assay (ELISA) for deter-mination of Staphylococcus aureus enterotoxin type B. HealthLab. Sci. 15:28-31.

25. OMS Serie de Rapports techniques. 1976. Aspects microbi-ologiques de l'hygiene des denrdes alimentaires. No. 598. Or-ganisation mondiale de la Sante, Geneva.

26. Peterkin, P. I., and A. N. Sharpe. 1984. Rapid enumeration ofStaphylococcus aureus in foods by direct demonstration ofenterotoxigenic colonies on membrane filters by enzyme immu-noassay. Appl. Environ. Microbiol. 47:1047-1053.

27. Porstmann, T., B. Porstmann, R. Wietschke, R. von Baehr, andE. Egger. 1985. Stabilization of the substrate reaction of horse-radish peroxidase with 0-phenylenediamine in the enzymeimmunoassay. J. Clin. Chem. Clin. Biochem. 23:41-44.

28. Rampling, M. W. 1974. The solubility of fibrinogen in solutionscontaining dextrans of various molecular weights. Biochem. J.143:767-769.

29. Salonen, E.-M., and A. Vaheri. 1981. Rapid solid-phase enzymeimmunoassay for antibodies to viruses and other microbes:effects of polyethylene glycol. J. Immunol. Methods 41:95-103.

30. Saunders, G. C., and M. L. Bartlett. 1977. Double-antibodysolid-phase enzyme immunoassay for the detection of staphy-lococcal enterotoxin A. Appl. Environ. Microbiol. 34:518-522.

31. Saunders, G. C., and E. H. Clinard. 1976. Rapid micromethod ofscreening for antibodies to disease agents using the indirectenzyme-labeled antibody test. J. Clin. Microbiol. 3:604-608.

32. Stiffier-Rosenberg, G., and H. Fey. 1978. Simple assay forstaphylococcal enterotoxins A, B, and C: modification of en-zyme-linked immunosorbent assay. J. Clin. Microbiol. 8:473-479.

33. Suter, M. 1982. A modified ELISA technique for anti-haptenantibodies. J. Immunol. Methods 53:103-108.

34. Thompson, N. E., M. Razdan, G. Kuntsmann, J. M. Aschen-bach, M. L. Evenson, and M. S. Bergdoll. 1986. Detection ofstaphylococcal enterotoxins by enzyme-linked immunosorbentassays and radioimmunoassays: comparison of monoclonal andpolyclonal antibody systems. Appl. Environ. Microbiol. 51:885-890.

35. Tijssen, P. 1985. Practice and theory of enzyme immunoassays,p. 297-328. In R. H. Burdon and P. H. van Knippenberg (ed.),Laboratory techniques in biochemistry and molecular biology,vol. 15. Elsevier, Amsterdam.

36. Todd, E. C. D. 1983. Foodborne disease in Canada-a 5-yearsummary. J. Food Prot. 46:650-657.

37. Wieneke, A. A., and R. J. Gilbert. 1985. The use of a sandwichELISA for the detection of staphylococcal enterotoxin A infoods from outbreaks of food poisoning. J. Hyg. 95:131-138.

38. Windemann, H., and E. Baumgartner. 1985. Application ofenzyme-linked immunosorbent assay (ELISA) with labeled an-tigen for detection of staphylococcal enterotoxins A, B and C infoods. Zentralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg.Abt. 1 Orig. Reihe B 181:320-344.

39. Windemann, H., and E. Baumgartner. 1985. Application ofsandwich-ELISA with labelled antibody for detection of staphy-lococcal enterotoxins A, B, C and D in foods. Zentralbl.Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. 1 Orig. Reihe B181:345-363.

APPL. ENVIRON. MICROBIOL.

on August 3, 2020 by guest

http://aem.asm

.org/D

ownloaded from