International Journal of Food Microbiology, 12 (1991) 91-102
© 1991 Elsevier Science Publishers B.V. 0168-1605/91/$03.50 FOOD 00367
Immunological methods for detection of foodborne pathogens and their toxins * S. N o t e r r n a n s a n d K. W e r n a r s Laboratory for Water- and Food Microbiology, National Institute of Public Health and Environmental Protection, Bilthooen, The Netherlands (Received 23 July 1990; accepted 24 September 1990)
Improved methods to detect microorganisms and their toxins introduced during the last decade involve among others recombinant DNA techniques and various immuno-assays such as the enzyme-linked immunosorbent assay and the latex agglutination. Immuno-assays are based on a quantitative reaction of an antigen (bacterial metabolite, e.g., toxin) with its antibody. Therefore, they are suited for detection of microorganisms based on their production of specific antigens and for quantitative detection of bacterial toxins. Sensitivity and specificity of immuno-assays are mainly determined by the antiserum used. In this respect the use of well selected monocional antibodies can be of advantage. With the enzyme-linked immunosorbent assay and latex agglutination test quantities of 0.1-1 ng of antigen/ml can be detected. Of both techniques the latex agglutination method has several advantages; the method is simple, inexpensive and rapid. Since each immuno-assay is sensitive to non-specific reactions, recognition of false positive results is necessary. The most appropriate method for this is to add an inhibitor to the test sample which blocks specifically the paratope of the immunoglobuline. Another general disadvantage of immuno-assays is that only the antigenicity is determined and this may differ from the actual toxicity. Therefore, antibodies should be used that react with the toxic centre(s) of the molecule, which can be accomplished by using well selected monoclonal antibodies.
Key words: Immunological methods; Pathogenic bacteria; Toxins
In judging the quality of food and foodproducts the microbial condition is of great importance and various laboratory methods have been developed to detect. enumerate and classify the organisms present. Conventional methods for the detection and identification of foodborne pathogenic microorganisms are laborious and time-consuming
(Salmonella, Listeria monocytogenes) o r r e q u i r e
Correspondence address: S. Notermans, Laboratory for Water- and Food Microbiology, National Institute of Public Health and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, The Netherlands. * The contents of this article have also been presented at the RAMI '90 conference held 7-10 June 1990, at Espoo, Finland.
92 laboratory (Clostridium botulinum, enterotoxigenic Staphylococcus aureus and Clostridium perfringens, diarrheogenic Escherichia coli). In order to circumvent these problems and yet be able to judge or predict the safety of food products, 'indicator organisms' are determined. The presence of high numbers of coliform organisms, for instance, may indicate a fecal contamination of the raw materials, whereas the presence of Gram-negative microorganisms may demonstrate insufficient pasteurization. High numbers of stapylococci may be a result of poor hygienic conditions during food processing. Detection of 'indicator organisms' is very simple: a suspension of the foodproduct is plated on a defined agar medium and incubated for 18-22 h under standard conditions, allowing rapid determination of a colony count. At present, manufacturers of food products are introducing rules of 'good manufacturing practice' and 'codes of production' to ensure the safety of these products. As a general rule, only raw ingredients are processed that are. free of pathogens. In addition, processing conditions are well defined, known to kill all microorganisms, and to avoid recontamination. Precise setting of intrinsic factors such as pH, wateractivity and redox potential and extrinsic factors such as gas packaging and storage temperature should prevent outgrowth of pathogens that might eventually be present in the final product. A general introduction of these rules and codes may drastically decrease the need for routine detection procedures of pathogens and their toxins in the near future. Testing for microorganisms will then be limited to specific critical control points in the production process. Beside microbial testing of critical control points there is a need to learn more about intrinsic and extrinsic safety factors in relation to newly recognized pathogens (Salmonella enteritidis, Listeria monocytogenes, Bacillus cereus) and of new food products (e.g. pasteurized meals without intrinsic safety factors). For these purposes and also for the field of epidemiology there is a need for rapid and reliable procedures for the detection and identification of foodborne pathogens and their toxins. Rapid methods, developed during the last decade have been reviewed by Huis in 't Veld et al. (1988) and include, among others, immuno-assays. An overview of immuno-assays and their principles for detection of foodborne pathogens and their toxins has recently been presented by Notermans (1989). This paper deals with the application of immuno-assays and their reliability.
Principles of immuno-assays Different immuno-assays have been developed during the last decade. Of the assays developed both the enzyme immuno-assay and the latex agglutination assay are at this moment the most widely applied assays. Enzyme immuno-assays can be divided into homogeneous and heterogeneous enzyme immuno-assays (Rubinstein et al., 1972). The latter includes competitive assays and non-competitive assays. Examples of both assays are summarized by Swaminathan and Konger (1986) and Fey (1983).
93 In the competitive assay, antigen in the test sample competes with enzyme-labelled antigen for binding sites on specific antibodies bound to a solid matrix. The binding of enzyme-linked antigen to immobilized antibody is decreased by the presence of antigen in the test sample. Therefore, the concentration of antigen is inversely proportional to the concentration of product formed by the enzyme reaction. Various modifications based on the principle of competition exist. The sandwich enzyme-linked immunosorbent assay (ELISA) is an example of a non-competitive assay. With this technique antibodies are coated onto a solid matrix. After incubation with the test sample, the amount of absorbed antigen is measured using enzyme-linked antibodies. The enzyme activity is now proportional to the concentration of the antigen. In assays like the sandwich ELISA, the antigen must contain at least two binding sites (epitopes) for the antibody. Of the immuno-assays recently developed the latex agglutination assay is the most simple method. The test is identical to the haemagglutination assay, however, the red blood cells are replaced by polystyrene latex particles (Salomon and Tew, 1968). The test procedure is as follows: latex particles are coated with antibodies. Test samples (food extracts, culture fluid, etc.) are added to the sensitized latex particles. After a short time of shaking the mixture is allowed to incubate for a certain period, mostly overnight. If antigen is present in the sample a network of sensitized latex particles with the antigen will occur. If no antigen is present no network will arise and the latex particles will form a sediment. Depending on the size of latex beads used and on the test procedure applied, agglutination may be made visible (Kamphuis et al., 1989). In general this last type of latex agglutination is more rapid and results are already obtained within 10 minutes.
Application of immuno-assays Immuno-assays can be used for detection of substances with immunological properties or substances which can be converted into immunologically active substances. These include proteins, glycoproteins, certain polysaccharides and haptens like mycotoxins. Therefore, not only bacterial protein toxins can be detected but also other metabolic products. Production of certain metabolic products is related to growth of the microorganism and quantification of these metabolites indicates the number of microorganisms present. Metabolites characteristic for a micro-organism or for a group of microorganisms can be used for typing of the organism, lmmuno-assays have been described for a large number of microbial toxins. Assays have been described, among others, for botulinum toxin (Notermans et al., 1979a), cholera toxin (Takeshi et al., 1983), and mycotoxins (Pestka et al., 1981; Fan et al., 1984). Immuno-assays for detecting microorganisms have been developed for e.g. Salmonella spp. (Swaninathan and Konger, 1986), Listeria spp (Farber and Speirs, 1987), Yersinia enterocolitica (Leirisalo et al., 1984) and moulds (Notermans et al., 1986; Lin et al., 1986).
Detection of bacterial toxins Staphylococcal enterotoxin (SE) Very small amounts of SE (less than 1 /~g) may have an emetic effect in man. Reiser et al. (1974) postulated therefore that for determining the safety of a given food, a detection method should be able to detect 0.125-0.250 ~g SE per 100 g of food. Therefore the most appropriate method for detecting SE in foods is the use of sensitive immuno-assays such as ELISA and latex agglutination. With these methods quantities of 0.1-1~tg of SE in 100 g of food can easily be recovered. Due to the sensitivity of these assays, a simple extraction procedure is sufficient to detect these small quantities. The detectable amount of recovered toxin (recovery) from a food product depends on the efficiency of the extraction procedure used and on the type of food. For extraction, foods are blended with equal amounts of a buffer solution (0.1 M phosphate buffer, pH 7.2, containing O.15 M NaCI). After centrifugation the supernatant can be tested for the presence of SE. Depending on the type of food, however, such a simple extraction is not always possible. In case of fat-containing extracts an additional chloroform extraction may be needed. If dry samples such as cereals have to be tested, more extraction buffer has to be used and for a sensitive detection of SE in such samples concentrating of the extract is necessary. Due to the simple extraction procedure, SE can be detected quantitatively by ELISA. However, one has to keep in mind that the recovery of SE depends on the food product tested. Therefore, for a quantitative detection of the toxin, recovery has to be tested as well. For this, uncontaminated food of the same type, to which varying quantities of purified SE have been added, may be suited (Sch~Snwalder et al., 1988). An additional problem in the SE detection is that other components, also present in food extracts, may yield false positive immunological reactions. For example, staphylococcal protein A present in test samples is known to cause false positive results (Koper et al., 1980). In testing extracts of canned mussels for the presence of SE, Essink et al. (1985) observed false positive results, probably caused by lysozyme. Botulinum toxin Confirmation of foodborne botulism is based on the detection and identification of the toxin in the blood serum of the patients as well as in the incriminated food. The quantities of toxin in blood serum are typically very low, whereas those in the incriminated food may be substantially higher. Botulinum toxins are extremely potent neurotoxins and therefore only ultrasensitive detection assays are of interest. These assays include the bioassay in mice and the highly sensitive immuno-assays like ELISA (Shone et al., 1985). The use of bioassays is common, but even so there are many drawbacks. A general disadvantage of the use of animals for determination of toxic substances like botulinum toxin is the great variation in results. Furthermore, the test is unsuitable for examination of samples containing other toxic substances that may cause interference or non-specific death in mice. Immuno-assays, however, suffer from other drawbacks. Antitoxins, such as polyclonal antibodies, react with immunological sites (epitopes) in the whole molecule
95 and may still react with the toxin even if it has been detoxified (Notermans and Nagel, 1989; Sugiyama, 1980). Furthermore the sensitivity of all in vitro immunological methods is less than that of the mouse intraperitoneal injection method. Even the ELISA technique described by Shone et al. (1986) using an amplifying technique is not more sensitive. The amplifying system is based on enzymatic transformation of NADP ÷ to N A D ÷ by alkaline phosphatase that is coupled to the IgG antibotulmum toxin. Each molecule of NAD ÷ produced is subsequently and rapidly cycled by certain enzymes that generate several hundreds of colored formazan molecules. Therefore the application of immuno-assay for detecting botulinum toxin is strongly limited. Clostridium perfringens enterotoxin (CPE) Several immuno-assays have been described for detection of CPE (Olsvik et al., 1982; Notermans et al., 1984). It is generally accepted that food poisoning caused by C. perfringens results from ingestion of large numbers of vegetative cells. Subsequent sporulation in the intestine results in release of CPE. CPE is only sporadically produced in food (Naik and Duncan, 1977; Craven et al., 1981). Thus, the potential health haTard posed by preformed CPE in food is of low significance. Preformed CPE will lose its activity largely during stomach passage. CPE released in the intestinal tract, however, maintains its immunological stability in stools over a long period of time as demonstrated by Notermans et al. (1984). Dowell et al. (1975) and Notermans et al. (1984) found that not all stools of patients attacked by C. perfringens contained CPE. They showed that the presence of CPE in stools of patients depends, among others, on the moment of sampling. To confirm a C. perfringens foodborne disease, by testing faeces for the presence of CPE by ELISA, it has to be taken into account that false negative results may be obtained. Faeces may contain certain products destroying the IgG coated on the polystyrene surface of the ELISA-trays (Notermans et al., 1984). Bacillus cereus enterotoxin (BCE) B. cereus is now well-established as a significant cause of foodborne illness in humans. The so-called 'diarrheal syndrome' usually associated with proteinaceous foods, vegetables, sauces and puddings is characterized by an incubation period ranging from 8-16 h. Symptoms generally resolve within 12-24 h. Goepfert et al. (1972) and Spira and Goepfert (1972) suggested the existence of an enterotoxin, following experiments involving the infection of whole-cell B. cereus cultures into ligated rabbit ileal intestinal loops. BCE may be preformed in food or produced in the small intestine after consumption of food contaminated by B. cereus (Kramer and Gilbert, 1989). However, up till now the significance of preformed toxin in the 'diarrheal syndrome' is not clear. Detection of pathogenic microorganisms
Detection of microorganismsis based on detection of characteristic metabolites. For the detection of pathogenic microorganisms the most appropriate metabolites
96 Table I Immuno-assays versus DNA-hybridization Ior detecting enterotoxigenic Staphylococcus aureus. Strains displaying a positive result in the immuno-assay were tested by DNA hybridization Immunological method
Enterotoxan(SE) tested SEA SEB
Agar gel diffusion ELISA Latex agglutination
4/4 a 4/4 8/5
7/7 16/14 48/3
7/7 18/15 3/2
a x / y ~ enterotoxin production as determined by immuno-assay: y enterotoxin gene present as de-
termined by hybridization are the toxins produced (C. botulinum, enterotoxigenic S. aureus, C. perfringens and E. coli). For organisms without a clarified pathogenic mechanism appropriate metabolites are mostly selected on the base of trial and error. For this purpose, culture fluid a n d / o r bacterial cells are injected into mice. Using the hybridoma cell technology, monoclonal antibodies are produced and selected for their suitability. The antigens reacting with these selected monoclonals are mostly extracellular substances released by the cell wall membrane and are often heat-stable. Staphylococcus aureus
Only pure cultures of S. aureus can be tested for enterotoxin production. Due to the sensitivity of the ELISA and latex agglutination a simplified enterotoxin production method is possible. Strains of S. aureus are inoculated in Erlenmeyer flasks containing brain-heart infusion (BHI)-broth. After incubation with shaking at 37 ° C for 48 h the culture fluid is tested for the presence of the enterotoxin. It has to be taken into account that especially in culture filtrates of S. aureus various metabolites may be present that are capable of causing non-specific reactions. We demonstrated such non-specific reactions by comparing immuno-assays and DNAhybridization assays for detecting enterotoxigenic S. aureus. Strains of S. aureus positive for enterotoxin production in different assays were tested for the presence of the gene encoding the toxin (Table I). Results obtained with the classical agar gel diffusion methods were identical to those obtained by D N A hybridization. Using the latter ELISA technique, 5 of 38 ELISA positive strains did not possess the gene encoding the toxin, and 49 of 59 latex agglutination-positive strains did not show hybridisation reactions. Clostridium perfingens
CPE is a characteristic metabolite of enterotoxigenic C. perfringens strains. Therefore this toxin can be used for detecting these organisms. CPE formation by C. perfringens is correlated with sporulation and in the past time consuming techniques have been described for testing strains (Stringer et al., 1982). We demonstrated (Notermans et al., 1984) that a simple one-step culturing technique may satisfy if the ELISA is used to detect the enterotoxin present. However, it became clear that a
97 strain can only be designated as a negative strain if enough spores are present in the culture fluid in the absence of CPE. In a comparative study (Van Damme-Jongsten et al., 1990) 38 strains of C. perfnngens were tested for production of CPE by ELISA. Twenty strains contained the gene encoding C P E as tested by D N A hybridization and 19 of these produced detectable quantities of CPE. One of the 18 strains lacking the C P E gene gave positive results in the ELISA. Clostridium botulinum To test the presence of C. botulinum in a sample the customary procedure is to enrich the sample in a suitable medium. After a proper incubation, culture supernatant is tested for the presence of toxin. The quantity of toxin produced depends on factors such as the type of sample, the presence of competitive microflora and incubation temperature (Notermans et al., 1979b). Generally, only small quanitities of toxin are produced. Although a number of different immuno-assays such as heamagglutination, latex agglutination, radio-immuno-assay and ELISA have been described for detecting botulinum toxin in culture supernantams, experience has made clear that these assays are less sensitive than the currently applied mouse-bioassay (Notermans, unpublished results). Escherichia coil Immunological assays using either polyclonal or monoclonal antibodies have been developed for identifying enterotoxigenic E. coli producing a heat-labile enterotoxin (LT). Recently we have tested about 800 E. coli strains for the presence of the gene encoding LT by DNA-hybridization. These 800 strains were randomly isolated from ground beef and pork and from poultry meat. N o n e of these strains contained the gene (Table II). Then 50 out of these 800 strains were additionally tested by using a
Table I1 Hybridization of Escherichia coli strains with a DNA probe encoding the heat-labile enterotoxin (LT) and production of LT tested by latex agglutination Origin of strains
Pork Beef Poultry meat Stools of patients with acute enteritis (170 stools tested) a
Hybridization Number of strains tested 274 248 278 8160
Hybridization reactions positive 0 0 0 157 (4 stools)
Latex agglutination Number of Agglutination strains reactions tested positive 8 0 18 0 14 0 20 with 20 LT gene 20 without 0 LT gene
a From each stool 48 random E. coil isolates were tested by DNA hybridization.
98 commercially available latex agglutination assay (Oxoid. Diagnostic Reagents, TD920). No positive reactions were observed. Immunological assays for detecting the heat-labile enterotoxin directly in stools are less sensitive in identifying E. coli infections (Morgan et al., 1983). Immuno-assays detect only those antigens in stools that survive enzymatic degradation, which can be considerable. Therefore E. coil present in stools should be tested for toxin production. We tested E. coli isolates from stools of patients with acute enteritis by D N A hybridization and latex agglutination (Table II). From each patient 48 isolates were tested. Of the 170 stools tested four contained LT-producing E. coll. The results obtained with latex agglutination tests of E. coli strains were completely in accordance with the D N A hybridization. Due to the low percentage of enterotoxigenic E. coli strains in food, routine screening of E. coil for toxin production by immuno-assay does not seem to be useful. The use of immuno-assay should be limited to foodborne disease outbreaks with an etiology resembling E. coli infection. Listeria, Salmonella, Campylobacter, Yersinia, etc To detect presence/absence of these pathogens in food, ELISA systems have been developed. However, these systems also suffer from the main disadvantage of having to culture the organisms before testing. This is necessary to produce sufficient antigen to allow a positive test result. Depending on the immunoglobulins used, ELISA can be used to differentiate e.g. invasive from non-invasive Campylobacterjejuni (Klipstein et al., 1985). From an efficiency study carried out by Beckers et al. (1988) it became clear that the baseline extinction values of ELISA systems depends on the type of food, which may lead to misinterpreation of the test results. Using reference samples (milkpowder containing small numbers of Salmonella) it was observed that false-negative results may be obtained by immunological methods as well as by culture-methods.
Reliability. of immuno-assay Antibodies used in the immuno-assay react with the immunological sites (epitopes) present on the molecule of interest. In case of a toxin, the antibodies may still react with the toxin even if it has lost its biological activity. It was demonstrated by Notermans (1989) that if botulinum toxin type B was added to surface water, at pH 8.1, and stored for several days at 20 ° C, biological activity was lost rapidly. This loss of biological activity, as demonstrated in a decrease of toxicity in mice, was not associated with a decrease in immunogenicity if polyclonal antibodies were used. However, if a monoclonal antibody, which neutralizes the toxin in vivo was used, the amount of determined toxin was equal to the amount of biologically active material present. Therefore, antibodies should be used which only react with the toxic site(s) of the molecule. This can be accomplished by using the monoclonal antibody technique (Shone et al., 1986; Kozaki et al., 1986). From a theoretical point of view, monoclonal antibodies which do not neutralize the toxin in vivo should not be used in an irnmuno-assay. If the epitope reactive with that mono-
99 clonal antibody is damaged, the toxin is not detectable anymore, while it cannot be excluded that the toxin is still biologically active. Each immuno-assay is potentially sensitive for non-specific reactions. These reactions occur if substances, other than the antigen, bind to the antibodies used in the assay. Especially in culture filtrates of bacteria, substances may be present which react in a non-specific way with immunoglobulines. An example of such a substance is protein A produced by S. aureus (Notermans et al., 1982). This protein A binds to the Fc-fragments of the IgGs used in the assay, giving rise to false positive reactions. Recently it was demonstrated by Tatini (personal communication), using the immunoblotting technique, that besides protein A a number of other cross-reacting substances are present in culture filtrates of S. aureus. These cross-reactions were observed with both polyclonal and monoclonal antibodies. Also, substances that give rise to false positive reactions may be present in food. It was reported by Essink et al. (1985) that lysozyme can strongly associate with immunoglobulins and form bridges between IgGs. In the latex agglutination assay for detecting moulds, Kamphuis et al. (1989) encountered false positive reactions when testing extracts of walnuts. These reactions were probably caused by tannins present in the extract. Besides false positive reactions, false negative reactions may be obtained. Latter reactions may occur if the antigen is masked by components present in the test sample. It was demonstrated by Schwabe et al. (1990) that, if heat-stable enterotoxin produced by S. aureus was added to an extract of peas and subsequently heated, toxin could not be detected by ELISA. However, after a high pH treatment of the extract, toxin was dectectable again.
Improvement of the reliability of immuno-assays Following on from discussion above, it has become clear that by using sensitive immuno-assays such as the ELISA and the latex agglutination, non-equivocal results may be obtained. Therefore, control experiments have to be introduced to check for both false negative and false positive results. False negative results can easily be recognized by addition of a known quantity of antigen to the test sample. For this purpose, internationally standardized antigen preparations would be preferred. It has to be emphasized that such standards are not yet available. Therefore, results obtained in different laboratories cannot be compared for reliability. For recognizing false postive results various possibilities exist. The most appropriate way is specific blocking of the IgG used in the assay by either synthetic epitopes or by anti-idiotype antibodies. Specific blocking is, however, only possible if the assay is carried out by using monospecific antibodies such as monoclonal antibodies. Kamphuis et al. (1989) described a latex agglutination for detecting moulds based on specific blocking of the monospecific antibodies used by synthesized epitopes. The monospecific antibodies for detecting P e n i c i l l i u m and Aspergillus species recognize the /3(1-5)-linked galactofuranose units of the extracellular polysaccharides (EPS) produced by these moulds. A synthetic tetramer of B(1-5)-linked
100 g a l a c t o f u r a n o s e specifically b l o c k e d the r e a c t i o n of the latex p a r t i c l e s w i t h the EPS of moulds. If a positive latex a g g l u t i n a t i o n is still positive after specific b l o c k i n g , the a g g l u t i n a t i o n has to be c o n s i d e r e d to be false positive. T h e n the s a m p l e in its c o m p o s i t i o n is n o t suitable for testing. T h e p o s s i b i l i t y of specific b l o c k i n g b y using an a n t i - i d i o t y p e a n t i b o d y was d e m o n s t r a t e d r e c e n t l y b y N o t e r m a n s et al. (1989). In their e x p e r i m e n t s an a n t i - i d i o t y p e a n t i b o d y was tested to recognize false positive E L I S A - r e a c t i o n s in d e t e c t i n g s t a p h y l o c o c c a l e n t e r o t o x i n B. F o r this p u r p o s e m o n o c l o n a l a n t i b o d i e s were used as c o a t i n g a n t i b o d y in the E L I S A . F ( a b ) 2 f r a g m e n t s of the a n t i - i d i o t y p i c m o n o c l o n a l a n t i b o d i e s were u s e d to b l o c k the p a r a t o p e s o f the c o a t e d a n t i b o d i e s specifically. In this w a y b i n d i n g o f a n t i g e n was p r e v e n t e d , w h i c h was d e m o n s t r a t e d b y the lack of b i n d i n g b y s u b s e q u e n t a d d e d e n z y m e - l i n k e d p o l y c l o n a l a n t i b o d i e s . A true positive r e a c t i o n can o n l y be c o n f i r m e d w h e n specific b l o c k i n g results in a n e g a t i v e .reaction in the E L I S A . F r o m these o b s e r v a t i o n s it b e c o m e s clear that the r e l i a b i l i t y of sensitive i m m u n o - a s s a y s can b e i m p r o v e d d r a s t i c a l l y in a s i m p l e way. H o w e v e r , p r o d u c t i o n of the specificly b l o c k i n g agent n e e d e d ( s y n t h e t i c e p i t o p e s o r a n t i - i d i o t y p i c a n t i b o d ies) is not always simple. T h e use o f i n t e r n a t i o n a l l y s t a n d a r d i z e d a n t i g e n s will i m p r o v e b o t h reliability of the assay a n d c o m p a r a b i l i t y .
References Beckers, H.J., Tips. P.D.. Soentoro, P.S.S., Delfgou-van Asch, E.H.M. and Peters, R. (1988) The efficacy of enzyme immuno-assays for the detection of salmonellas. Food Microbiol 5, 147-156. Craven. S.E., Blankenship. L.C. and McDonel, J.L. (1981) Relationship of sporulation, enterotoxin formation, and spoilage during growth of Clostridiura perfringens type A in cooked chicken. Appl. Environ. Microbiol. 4, 1184-1191. Dowell, Jr., V.R.. Torres-Anjel, M.J.. Riemann, H.P., Merson, M,, Whaley, D. and Darland, G. (1975) A new criterion for implicating Clostridium perfringens as the cause of food poisoning. Rev. Lat. Am. Microbiol. 17, 137-142. Essink, A.W.G.. Arkesteijn. G.J.M.W. and Notermans, S. (1985) Interference of lysozyme in the sandwich enzyme-linked immunosorbent assay (ELISA). J. Immunol. Methods 80, 91-96. Fan, T.S.L.. Zhang, G.S. and Chu. F.S. (1984) Production and characterization of antibody against aflatoxin Q1- Appl. Environ. Microbiol. 47, 526-532. Farber. J.M. and Speirs, J.I. (1987) Monoclonal antibodies directed against the flagellar antigens of Listeria species and their potential in EIA-based methods. J. Food Protect. 50, 479-484. Fey, H. (1983) Nachweis yon Staph.vlokokken Enterotoxinen in Lebensmitteln. Schweizerische Gesellschaft fiir Lebensmittelhygiene. Schriftenreihe Heft 13, 51-67. Goepfert. J.M., Spira. W.M. and Kim, H.U. (1972) Bacillus cereus: a food poisoning organism. J. Milk Food Technol. 35. 213-227. Huis in 't Veld. J., Hartog, B. and Hofstra, H. (1988) Changing perspectives in food microbiology: Implementation of rapid rmcrobiological analyses in modem food processing. Food Rev. Int. 4. 271-329. Kamphuis, H.J., Notermans, S., Veeneman, G.H., van Boom, J.H., and Rombouts, F.M. (t989) A rapid and reliable method for detection of molds in food: Using the latex agglutination assay. J. Food Protect. 52. 242,-247. Klipstein, F.A., Engert. R.F.. Short. H. and Schenk. E.A. (1985) Pathogenic properties of Campylohacter jejuni: Assay and correlation with clinical manifestations. Infect. Immun. 50, 43-49.
101 Koper, J.W., Hagenaars, A.M. and Notermans. S. (1980) Prevention of cross-reactions in the enzymelinked immunosorbent assay (ELISA) for the detection of Staphylococcus aureus enterotoxin type B in culture filtrates and foods. J. Food Safety 2. 35-45. Kozald, S., Kamata, Y., Nagai, T., Ogasawara. J. and Sakagucki, G. (1986) The use of monoclonal antibodies to analyze the structure of Clostridium botulinum type E derivative toxin. Infect. Immun. 52, 786-791. Kramer, J.M. and Gilbert, R.J. (1989) Bacillus cereus and other Bacillus species. In: M.P. Doyle (Ed.), Foodbome Bacterial Pathogens. Marcel Dekker, New York, NY. and Basel. pp. 21-70. Leirisalo, M., Gripenberg, M., Julkunen, 1. and Repo. H. (1984) Circulating immune complexes in Yersinia fection. J. Rheumatol. 11, 365-368. Lin. H.H., Lister, R.M. and Cousin, M.A. (1986) Enzyme-linked immunosorbent assay for detection of mold in tomato puree. J. Food Sci. 51, 180-192. Morgan, D.R., DuPont, H.L., Wood, L.V. and Ericsson, C.D. (1983) Comparison of methods to detect Escherichia coli heat-labile enterotoxin in stool and cell-frec culture supernatants. J. Clin. Microbiol. 18, 798-802. N a i l H.S. and Duncan. C.L. (1977) Enterotoxin formation in food by Clostridium perfringens t)~e A. J. Food Safety 1, 7-18. Notermans, S. (1989) Detection of toxigenic microorganisms and their toxins by newer methods. Proc. Int. Seminar, Santander, pp. 323-336. Noterrnans, S. and Nagek J. (1989) Assays for botulinum and tetanus toxins. In: L.L. Simpson (Ed.), Botulinum Neurotoxin and Tetanus Toxin. Academic Press. London, pp. 319-331. Notermans, S., Dufrenne, J. and Kozaki, S. (1979a) Enzyme-linked immunosorbent assay for detection of Clostridium botulinum type E toxin. Appl. Environ. Microbiol. 37. 1173-1176. Notermans, S., Dufrenne, J. and van Schothorst, M. (1979b) Recovery of Clostridium botulinum from mud samples incubated at different temperatures. Eur. J. Appl. Microbiol. Biotechnol. 6, 403-407. Notermans, S., Timmermans, P. and Nagel, J. (1982) Interaction of staphylococcal protein A in enzyme-linked immunosorbent assays for detecting staphylococcal antigens. J lmmunol Methods 55. 35-41. Notermans, S., Boot. R., Tips, P.D. and de Nooij, M.P. (1983) Extraction of staphylococcal enterotoxins (SE) from minced meat and subsequent detection of SE with enzyme-linked immunosorbent assay (ELISA). J. Food Protect. 46, 238-241. Notermans, S., Heuvelman. C., Beckers, H. and Uemura, T. (1984) Evaluation of the ELISA as tool in diagnosing Clostridmra perfringens enterotoxins. Zbl. Bakt. Hyg., I Abt, Orig. B 179. 225-234. Notermans. S., Heuvelman, C.J.. van Egmond. H.R., Paulsch, W.E. and Besling, J.R. (1986) Detection of mold in food by enzyme-linked immunosorbent assay. J. Food Protect. 49, 786-791. Notermans, S., Alber, G.. Sailer-Kramer, B. and Hammer, D.K. (1989) The use of an anti idiotype monoclonal antibody for reliable detection of staphylococcal enterotoxins. Food Agric. lmmunol. 1, 11-17. Olsvik, O., Granum. P.E. and Berdal. B.P. (1982) Detection of Clostridium perfringens type A enterotoxin by ELISA. Acta Pathol. Microbiol, lmmunol, Scand. Sect. B 90. 445-447. Pestka, J.J.. Steinert, B.W. and Chu, F.S. (1981) Enzyme-linked immunosorbent assay for detection of ochratoxin A. Appl. Environ. Microbiol, 41, 1472-1474. Reiser, R., Conaway. D. and Bergdoll, M.S. (1974) Detection of staphylococcal enterotoxin in food. Appl. Microbiol. 27.83-85. Rubenstein, K.E.. Schneider, R.S. and Ullman, E.F. (1972) "Homogeneous" enzyme immuno-assay. A new immunochemical technique. Biochem. Biophys. Res. Commun. 47. 846-851. Salomon, L.L. and Tew. R.W. (1968) Assay of staphylococcal enterotoxin B by latex agglutination. Pro