Letters in Applied Microbiology 1986, 2, 53-56

Evaluation of an enzyme immunoassay technique for the detection of salmonellas in minced meat

H.J. B E C K E R S , P.D. TIPS, E . D E L F O O U - V A N A S C H & R. PETERS Oiboratoryjor Water and Food Microbiology, National Institute of Public Health and Environmental Hygiene, P.O. Box I , 3720 B A Bilfhoven, The Netherlands

Received 12 December I985 and accepted 18 December 1985

BBCKBKS, H . J . , T I P S , P.D., D E L F G O U - V A N A S C H , E . & PBTEKS, R. 1986. Evaluation of an enzyme immunoassay technique for the detection of salmonellas in minced meat. Letters in Applied Miclohioloyy 2, 53-56.

A comparison was made between a commercially available enzyme immunoassay (ELISA) and various cultural procedures for detecting salmonellas in minced meat contaminated with a standard inoculum. To detect salmonellas by the ELISA tech- nique, it was necessary to modify the recommended baseline for spectrophotometric measurements to avoid false-positive results. The incidence of false-negatives was no greater than that obtained with a standard isolation procedure. Both methods were affected by the competing microflora

Several methods have been developed to reduce the time necessary for the detection of salmon- ellas in foods. Recently an enzyme immunoassay (ELISA) with a monoclonal antibody has been described which detects a salmonella flagellar antigen (Robison et ul. 1983). This procedure has been refined by Mattingly and Gehle (1984) and is now commercially available as the Sal- monella Bio-Enzabead Test Kit (Litton Bio- netics, Inc., Laboratory Products Division, Charleston. SC 29405, USA). The present study was carried out to evaluate the kit and to compare the results obtained with those from standard and other cultural procedures for the detection of salmonellas. For this purpose a standardized Salmonella inoculum (Beckers et a/ . 1985) was added to minced meat.

Materials and Methods

S A M P L E S

Standardized inocula consisted of 0.2 g of spray-dried milk, artificially contaminated with nalidixic acid-resistant Salmonella typhimurium I1 505 contained in gelatin capsules. The mean contamination level was five salmonellas/

capsule. Growth patterns of the nalidixic acid-resistant S. typhirnurium in salmonella enrichment media did not differ from those of the non-resistant strain, normally present in the reference material.

A sample of 1 kg of minced meat was col- lected from a local butcher’s shop and tested before use for the presence of salmonellas (Anon. 1981); they could not be detected in 25 g.

M E D I A

The following media were used: buffered peptone water (BPW) prepared according to ISO-6579 (Anon. 1981); Rappaport-Vassiliadis magnesium chloride-malachite green broth (RV) prepared according to Vassiliadis (1983); phenol red brilliant green agar (BGA) prepared accord- ing to ISO-6579 (Anon. 1981); violet-red bile agar (Difco 0012) supplemented with 10 g/l of glucose and 40 mg/l of nalidixic acid (VRBG + Nal); ENDO agar prepared from individual ingredients (ENDO).

For the enzyme immunoassay the Salmonella Bio-Enzabead Test Kit was kindly supplied by Litton Bionetics, Inc., Laboratory Products Division, Charleston, SC 29405, USA.

54 H . J . Beckers et al. P R O C E D U R E

One salmonella capsule was added to each of 25 x 100 ml of BPW. In parallel, 1 kg of minced meat was mixed with 9 1 of BPW and divided into 100 lots of 10 g; 25 of these were left unin- oculated whilst a further 25 samples were inocu- lated with two, four and eight capsules respectively. All samples were incubated at 37°C (see Fig. 1). After 18 h 0.1 ml of each BPW culture was transferred to a tube containing 10 ml of RV and incubated at 43°C. After 24 h a loopful of the RV culture was streaked onto BGA for the isolation of salmonellas according to a standard procedure (Anon. 1981). At the same time, 0.1 rnl of the culture in RV and of a decimal dilution in peptone saline was spread on VRBG + Nal to count the nalidixic acid- resistant Salmonella and on ENDO to count ‘total’ Enterobacteriaceae. Furthermore, 1 ml of each culture in RV was transferred to a tube containing 9 ml of BPW and incubated at 37°C in a water bath. After 6 h, a loopful of this culture was streaked on VRBG + Nal. Subse- quently, the culture was centrifuged at 5000 g for 20 min. The cell pellet was resuspended in 1 ml of phosphate buffered saline (PBS, pH 7.2) and heated for 20 min at 100°C in a boiling water bath. The heat-treated cell suspension thus obtained was the antigen used in the enzyme immunoassay. The principle of the sal- monella ELISA has been described by Emswiler-Rose et al. (1984). The test was carried out according to the manufacturer’s instruc-

Pre-enrichment in BPW at 37’C

after I+ih Selective enrichment in RV ot 4 S C

after 24 h

I I Isolation on BGA Enumeration of salmonelbs on VRBG + Nal Enumeration of Enterobocteriaceae on ENDO

Enrichment in BPW ot 37°C after 6 h

L ; , A Isolation on VRBG + Nol

t oft= 48 h

Isolation on BGA Isolation on VRBG + No 1

L Fig. 1. Procedure for the evaluation of the ELISA for the detection of salmonellas.

tions. Extinctions were measured at 405 nm with a Titertek Multiskan spectrophotometer. The remainder of each RV culture was incu- bated at 43°C for another 24 h and subcultured again on BGA according to the standard pro- cedure (Anon. 1981), as well as VRBG + Nal.

Results

Four negative controls for the ELISA supplied by Litton Bionetics, Inc., gave a mean extinction (Eneg) value of 0.24 at 405 nm (s,neg = 0.03). The 25 samples of minced meat, which had not been inoculated with the Salmonella, gave a mean extinction (E,,,,,) value of 0.33 at 405 nm (s,blank = 0.03). With the latter samples, the highest extinction was 0.40. Since the minced meat samples were considered to be Saltnonellu- free, it was decided to consider samples positive when reading > Eblank + 3 x s,blank ( = 0.42). The interpretation of the enzyme immunoassays utilized this baseline.

Results are presented in Table 1. Twenty-one capsules out of 25 gave positive Salmonella iso- lations after 24 h of selective enrichment in RV. The mean number of salmonellas (loglo) was 6.2. Further incubation, either in RV or, after subculture, in BPW resulted in one more posi- tive isolation. Cultural methods and the ELISA scored equally. This result was in accordance with the calculated probability of a negative capsule. Salmonellas could not be detected in 25 samples of the uninoculated minced meat, either by cultural methods or by the ELISA. This result demonstrated that the ELISA did not give false-positives. Inoculation of 25 samples of minced meat with two capsules each resulted in a mean number of salmonellas (log,,) of 3.3 in five selective enrichments, but the standard cul- tural method, as well as the ELISA, failed to detect the organism. Nevertheless, salmonellas could be detected in nine cultures in BPW when competing micro-organisms were suppressed on VRBG + Nal. With a similar suppression, sal- monellas could be detected in six samples after prolonged incubation of the culture in RV, whereas only one sample was found positive by the standard procedure. Similar results were obtained with samples of minced meat inocu- lated respectively with four and eight capsules. Salmonellas could be counted in 13 and 22 samples respectively, while the standard cultural procedure detected only two and three positive

Table 1. Detection of salmonellas in 25 capsules, in 25 samples of minced meat without capsules and in increasing numbers of capsules by the procedure shown in Fig. 1

After 24 h selective After 48 h selective enrichment in RV 6 h in BPW enrichment in RV

After a further

Count on VRBG + Nal:

numbers of Isolation salmonellas* Count on ELISA: Isolation on Isolation Isolation on on BGA: (numbers of ENDO: numbers VRBG + Nal: on BGA: VRBG + Nal:

numbers of relevant numbers of of numbers of numbers of numbers of positive positive Enterobac- positive positive positive positive

Substrate isolations samples) teriaceae* samples isolations isolations isolations

5.

e 3

Capsules

Minced meat alone 21 6.2 f 1.3(21) 6.8 0.9 22 22 22 22

with 0 capsules 0 <2.0 (0) 7.5 & 0.8 0 0 0 0 with 2 capsules 0 3.3 + 2.6 (5) 7.3 0.7 0 9 1 6 with 4 capsules 2 3.9 2.3(13) 7.2 + 0.4 9 21 8 14 with 8 capsules 3 4.8 i 1.5(22) 7.2 1.2 11 24 10 23

m. 3

s. 3 m m R

* In '"log-units, with 95% limits

56 H . J . Beckers et al. samples and the ELISA 9 and 11 . At the same time, however, salmonellas could be detected in 21 and 24 cultures in BPW, preceding the ELISA. The numbers of positive Salmonella iso- lations increased upon prolonged incubation of RV. After 48 h of incubation the standard cul- tural procedure detected only one positive sample less than the ELISA, but isolation on VRBG + Nal demonstrated that both methods failed to detect salmonellas in a large number of cases.

Discussion

According to the manufacturer, a valid result in the ELISA test requires the mean extinction of the negative controls to be < 0.12 at 405 nm. In the present study, this value was 0.24. No expla- nation could be offered for this striking differ- ence. Again, according to Litton Bionetics Inc., values > 0.20 at 405 nm could be considered positive. Minced meat samples, which were apparently Salmonella-free, gave extinctions up to 0.40. O n such a basis, either the tests were invalid, or all the results obtained were positive. However, previous experience with enzyme immunoassays in our laboratory (Notermans 1984) suggests that the ELISA test is entirely valid and another yardstick for a positive result was established on the basis of results obtained with negative minced meat samples.

Emswiler-Rose et a/. (1984) found complete agreement between ELISA and cultural methods when used in the examination of poultry and pork products. The good per- formance of the ELISA is confirmed by the results of the present study in which ELISA detected one more positive sample than the standard cultural procedure. For the latter result RV had to be incubated for up to 48 h. Despite the high numbers of ‘total’ Enterobacte- riaceae present, ELISA gave no false-positive results but false-negatives were obtained on some occasions because salmonellas were found to be present in corresponding enrichment cul- tures in relatively high numbers. An explanation for this might be the masking of antibodies by other Enterobacteriaceae. However, the effect is not dependent upon the number of Enterobac- teriaceae present, because this was high in both

positive and negative ELISA results. Instead, i t might depend on the type(s) of Enterobacte- riaceae. With both ELISA and the standard cul- tural method results were affected by the competing flora. Salmonellas could be detected in many more enrichments where this flora was suppressed by nalidixic acid at the stage of iso- lation.

A clear requirement was the need to prolong incubation of RV to 48 h in order to obtain a result with the standard cultural method which was equivalent to that from the ELISA. Accord- ing to the original description of the RV method (Vassiliadis 1983), an incubation time of 24 h would be sufficient. On the basis of our results we would recommend that RV is incubated for 48 h and streaked on the isolation medium after both 24 h and, in case of negative results, after 48 h.

The authors wish to thank Dr G.C. Mead, Food Research Institute, Bristol, UK for his critical review and correcting of the manuscript.

References

ANON. 1981 International Organization of Stan- dardization. Microbiology-General guidance for the detection of Salmonella. ISO-6579.

BECKERS, H.J., VAN LEUSDEN, F.M., MEYSSEN, M.J.M. & KAMPELMACHER, E.H. 1985 Reference material for the evaluation of the standard method for the detection of salmonellas in foods and feeding stuffs. Journal of Applied Bacteriology 59, 507-512.

EMSWILER-ROSE, B., GEHLE, W.D., JOHNSTON, R.W., OKREND, A,, MORAN, A. & BENNETT, B. 1984 An enzyme immunoassay technique for detection of salmonellae in meat and poultry products. Journal of Food Science 49, 1018-1020.

MATTINGLY, J.A. & GEHLE, W.D. 1984 An improved enzyme immunoassay for the detection of Salmon- ella. Journal of Food Science 49, 807-809.

NOTERMANS, S.H.W. 1984 Enzyme-linked immuno- sorbent assay of staphylococcal enterotoxins in foods. In Rapid Methods and Automation in Micro- biology and Immunology ed. Habermehl, K.O. pp. 649-655. Berlin : Springer Verlag.

ROBISON, B.J., PRETZMAN, C.I. & MATTINGLY, J.A. 1983 Enzyme immunoassay in which a myeloma protein is used for detection of salmonellae. Applied and Environmental Microbiology 45, 18 16-1 821.

VASSILIADIS, P. 1983 The Rappaport-Vassiliadis (RV) enrichment medium for the isolation of salmonellas: An overview. Journal of Applied Bacteriology 54, 69-76.