APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 1977, p. 518-522 Copyright © 1977 American Society for Microbiology

Vol. 34, No.5 Printed in U.S.A.

Double-Antibody Solid-Phase Enzyme Immunoassay for the Detection of Staphylococcal Enterotoxin A GEORGE C. SAUNDERS* AND MARY L. BARTLETT

Los Alamos Scientific Laboratory of the University of California, Los Alamos, New Mexico 87545

Received for publication 15 February 1977

A simple double-antibody enzyme immunoassay that uses a microtechnique developed for detecting staphylococcal enterotoxin A in food products. Sample preparation can be completed in less than 15 min. Assay sensitivity ranges from 0.4 ng (20-h test time) to 3.2 ng (1- to 3-h test time)4Of toxin per ml of prepared sample. Separation and detection of enterotoxin from spiked food products ranged between 72 and 98% of the amount added.

was

We pursued development of an enzyme immunoassay (EIA) for staphylococcal enterotoxin A (SEA) detection in foods because present methods for sample preparation and detection of enterotoxins are laborious and time consuming (2, 5, 7, 9, 12, 16). A desirable test for SEA detection should require relatively simple sample preparation and an ability to detect 125 to 250 ng of SEA per 100 g of food (9). Our results indicate that EIA may meet these requirements. The extraction from food and quantitative detection of SEA takes less than 3 h (sensitivity of 1.6 to 3.2 ng/ml) as compared with days for many other procedures (2, 9, 16). Enzyme immunoassays for detecting both antibodies and antigens (4, 10, 11, 13-15) are, with good reason, becoming increasingly popular. EIA offers sensitivity in the same range as radioimmunoassay (RIA) but without the problems of laboratory safety, disposal, short reagent shelf life, and equipment expense inherent to RIA procedures. MATERIALS AND METHODS Staphylococcal enterotoxin A. Purified staphylococcal enterotoxin A (lot CA-2-E, USA-BDRC) was supplied by H. G. Fugate, U.S. Department of Agriculture, Beltsville, Md. Antibody. Rabbit antiserum to staphylococcal enterotoxin A (lot FDA-69-36) was supplied by H. G. Fugate. The gamma-globulin fraction of the antiserum was obtained by ammonium sulfate precipitation (6). Conjugated antibody. Horseradish peroxidase (type VI, Sigma Chemical Co., St. Louis, Mo.) was conjugated to fractionated gamma-globulin at a ratio of 1.5 molecules of enzyme to 1 molecule of gammaglobulin by the method of Nakane and Kawaoi (8). The conjugate was filtered and stored until used at 4°C in glass without preservatives (plastic slowly inactivates the conjugate). Microplate preparation. Disposable microplates (type IS-FB-96, Linbro Chemical Co., New Haven, Conn.) were used as test vehicles. To assure firm

binding of anti-staphylococcal gamma-globulin (ASG), microplate wells were pretreated with 0.05 ml of a bovine serum albumin (BSA) aqueous solution (0.2 mg of BSA per ml). After air drying, the BSA was fixed to the plates with glutaraldehyde as previously described (11). Then 0.05 ml of a 1:200 dilution of unlabeled ASG was added to each well and was dried rapidly by blowing air at 45°C. Other dilutions of ASG were used in preliminary experiments, but 1:200 proved best for our purposes. Prepared microplates were stored in a dry, clean environment at room temperature. Trays retain their activity for at least 1 month. Just before use, the trays were rinsed twice in 0.15 M saline. Food product preparation. (i) Whole milk. Whole milk was brought to a pH of 8.0 by dropwise addition of 1% NaOH. Tween 80 (Fisher Scientific, Fair Lawn, N.J.) was added to a final volume of 0.25%. (ii) Mayonnaise. One milliliter of normal saline containing 0.5% Tween 80 was added per gram of mayonnaise sample. The mixture was homogenized for 3 min at the intermediate setting (-250 rpm) of a Virtex 23 Homogenizer (Virtex, Gardiner, N.Y.). The homogenate was brought to pH 8.0 by dropwise addition of 1% NaOH. (iii) Hot dog. One milliliter of normal saline containing 0.5% Tween 80 was added per gram of hot dog sample. The mixture was ground for 3 min with a porcelain mortar and pestle. The macerate was then squeezed through a gauze pad. The filtrate was brought to pH 8.0 by dropwise addition of 1% NaOH. EIA procedure. Binding of various known amounts of SEA in food extracts to adsorbed antibody was determined as follows. A 1-ml sample of the prepared extract was spiked with 100 ng of SEA. Then double dilutions (in food extract) through at least seven tubes were prepared. A tube containing food extract but no SEA was also prepared. Triplicate aliquots, 0.05 ml, of each dilution and the blank were placed into wells of an ASG-sensitized microplate. The tray was sealed with plastic tape, and the samples were incubated for 1 h at room temperature. The trays were then washed eight times in a stream of wash fluid containing phosphate-buffered saline, pH 7.8, and 0.05% Tween 80. After shaking out as much 518

EIA FOR SEA

VOL. 34, 1977 residual wash fluid as possible, 0.05 ml of a 1:100 dilution (in 0.5 M NaCl-1% Tween 80, pH 7.8) of ASG conjugate was added to each well. After again sealing the tray, it was incubated for 1 h at room temperature and then washed as before. Then 0.1 ml of a substrate solution consisting of 0.4 mM ABTS [2,2'-azino-di(3(Boehringer ethyl-benzthiazoline-6-sulfonate)], Mannheim Corp., Indianapolis, Ind.) in 0.1 M citratephosphate buffer, pH 4.0, and 1.3 mM H202 was added to each well. After a 30-min incubation at room temperature, the substrate solution was transferred with a Pasteur pipette into glass tubes containing 1 ml of 0.1 M citrate-phosphate buffer, pH 2.8. Absorbance of each diluted sample at 414 nm was determined in a Beckman model DU spectrophotometer with 1-cm quartz cuvettes. From this value, absorbance of the undiluted reaction product was calculated and tabulated. Averages of the replicates were determined. Standard curves were generated by plotting the data on semilog graph paper. Effects on test sensitivity by both increasing and decreasing the above incubation times were also obtained. Several 30-g samples of mayonnaise and hot dog extract were also spiked with known amounts of SEA and processed to determine whether SEA in situ could be recovered by the relatively simple extraction and test procedures described above. Absorbance values obtained were plotted against the standard curve, and dilution factors were applied to assay the amount of SEA separated from the food.

519

conjugate is incubated for 2 h followed by substrate for 15 min. Table 3 indicates that good sensitivity (3.2 ng/ml) is still obtained when SEA-contaminated hot dog extract and ASG conjugate incubation times are reduced to 15 min each; substrate reaction time was 30 min. This shorter procedure yields a curve with a reduced slope, but it may be useful as a screening test. The ability to accurately assay SEA in spiked mayonnaise and hot dog extract is illustrated in Table 4. Food containing as little as 2.5 ng of added enterotoxin per g yielded at least 70% of the enterotoxin after extraction. DISCUSSION

The purpose of this report has been to present a simple and sensitive method for the detection of SEA. This procedure should be useful to those workers who have a need to rapidly and accurately detect SEA concentrations in various organic materials. Presence of large amounts of organic matter in food extract test samples did not appear to interfere with the EIA test. This important finding, coupled with the test's high degree of sensitivity, should eliminate the elaborate purification and concentration steps currently required RESULTS in other serological procedures for SEA detecThe ability to detect known quantities of SEA tion in these materiais. in food extracts by EIA, using various reaction Because of budget limitations and other comtimes, is illustrated in Tables 1 through 3 and mitments, we have not been able to evaluate in Fig. 1, which illustrates the detectability of the EIA procedure for detecting levels of SEA SEA in milk (3.2 ng/ml), mayonnaise (1.6 in naturally contaminated products. However, ng/ml), and hot dog extracts (1.6 ng/ml), utiliz- we see no fundamental reason why this proceing incubation times of 1 h for enterotoxin, 1 h dure should not yield favorable results in this for ASG conjugate, and 30 min for substrate. area. Certain foods (for example some types of With hot dog extract as an example, Table 1 cheese) may contain proteolytic enzymes that shows the reproducibility of the data presented could conceivably attack the bound ASG, in Fig. 1. Table 2 shows the increase in sensitivity thereby interfering with assay sensitivity. In (0.4 ng/ml) obtained when hot dog extract con- these instances, such enzymes could be either taining SEA is incubated for 18 h and ASG heat or chemically inactivated (for example TABLE 1. Detection of SEA in hot dog extracta Absorbance at 414 nm SEA concn (ng/ml)

None (blank)Y 1.6 3.1 6.3 12.5 25 50 100

Reaction times: With standard deviation indicated. c Hot dog extract only. d p < 0.05 as compared with blank.

a

b

--Sample/blank

Avg" Replicate 3 Replicate 2 Replicate 1 0.190 0.170 0.200 0.190 ± 0.02 0.370d ± 0.02 0.370 0.380 0.350 0.470 ±0.06 0.520 0.500 0.400 0.690 + 0.06 0.660 0.650 0.750 1.07 ±0.08 1.06 1.15 1.00 1.62 ± 0.07 1.57 1.60 1.70 2.40 ±0.09 2.42 2.48 2.30 2.97 ± 0.07 2.94 3.05 2.92 60 min for enterotoxin, 60 min for conjugate, 30 min for substrate.

1.95 2.47 3.63 5.63 8.53 12.63 15.63

520

SAUNDERS AND BARTLETT

APPL. ENVIRON. MICROBIOL.

TABLE 2. Increase in assay sensitivity by lengthening enterotoxin and conjugate incubation timesa Absorbance at 414 nrm SEA concn (ng/ml)

Sample/blank

Avg" 0.100 0.150 0.130 ± 0.03 None (blank)c 0.1 0.140 0.160 0.150 0.150 ± 0.01 1.15 0.190 0.220 0.210 0.210 ± 0.02 0.2 1.62 0.250 0.250 0.250 0.250d 0.4 1.92 0.340 0.250 0.300 0.300 ± 0.05 0.8 2.31 1.6 0.350 0.330 0.390 0.360 ± 0.03 2.77 0.460 0.440 ± 0.03 3.1 0.410 0.440 3.38 0.730 0.700 0.750 6.3 0.730 ± 0.03 5.62 1.18 1.38 1.24 1.27 ± 0.10 12.5 9.77 2.57 2.42 2.52 2.50 ± 0.08 25 19.20 3.29 3.42 3.37 3.36 ± 0.07 25.90 50 3.46 4.06 3.77 ± 0.30 100 3.80 29.00 a Enterotoxin diluted in hot dog extract with 0.01% NaN3 added to prevent microbial growth. Reaction times: 18 h for extract, 2 h for conjugate, 15 min for substrate. I With standard deviation indicated. Hot dog extract only. d p < 0.05 compared with blank. Replicate 1

Replicate 2

Replicate 3 0.130

c

TABLE 3. Effect of reducing enterotoxin and conjugate incubation timesa Absorbance at 414 rum SEA concn (ng/ml)

Replicate 1

Replicate 2

Replicate 3

Sample/blank Avgh 0.090 ± 0.02

0.110 0.080 0.080 None (blank)Y 0.120 0.120 0.8 0.120 0.120 1.33 0.180 0.160 0.130 0.160 ± 0.03 1.6 1.78 0.240 0.200 0.190 3.1 2.33 0.210d 0.03 0.200 0.240 0.250 0.230 ± 0.03 2.56 6.3 0.300 0.300 0.300 ± 0.01 3.33 12.5 0.310 0.420 0.440 0.420 ± 0.02 25 0.410 4.67 0.530 0.530 50 0.470 0.510 ± 0.03 5.67 100 0.690 0.660 0.570 0.640 ± 0.06 7.11 a Enterotoxin diluted in hot dog extract. Reaction times: 15 min for extract, 15 min for conjugate, 30 min for substrate. b With standard deviation indicated. I Hot dog extract only. d p c 0.05 compared with blank.

chemicals which mercurate sulfhydryl groups) without significantly altering the antigenicity of the enterotoxin. It is also possible that the variance in background (the value obtained for an SEA-free sample) among various lots of a similar foodstuff could cause some decrease in sensitivity if one depended on the average of all lots to establish background values. Should this become a problem, one could, without much difficulty, include abbreviated standard curves for each lot of foodstuff tested. For example, in a 96-well microtiter tray, allowing 3 wells per point, one could use 18 wells to generate a 6point (0 to 25 ng/g) standard curve. The remaining 78 wells could be used to screen 26 suspect samples for the same lot. Enterotoxin A was selected for these experiments because of its reported difficulty of recovery (9) and also because it is the type most

commonly involved in staphylococcal food poisoning (1). We did not observe any difficulty in accurately quantitating enterotoxin artificially introduced into either mayonnaise or hot dogs. The data obtained from double-dilution experiments indicate that EIA approximates the required sensitivity of being able to detect 125 to 250 ng of enterotoxin per 100 g of product, and also has at least equivalent sensitivity to that of reported RIA procedures (3, 7). To detect minimal quantities of enterotoxin from foods, and for more exact quantitation of amounts detected in the relatively flat part of the standard curve, the toxin concentration obtained in the food extract could be increased by pervaporation or molecular filtration techniques. However, the increased sensitivity of EIA by use of the longer incubation protocol should make this type of procedure unnecessary for most applications.

VOL. 34, 1977

EIA FOR SEA

521

blank ratio of this point of significance ranged from 1.9 to 2.3. Previous reports from this laboratory (10, 11), reporting on enzyme-labeled antibody procedures, utilized the chromogen 5 aminosalicylic acid hydrochloride (5AS). We have changed to ABTS because it is commercially available in a pure form. Also, on a molar basis, it is considerably more sensitive than 5AS, and stock solutions appear to be stable indefinitely. Details of the optimization of ABTS for EIA procedures will be published elsewhere. Pretreatment of trays with BSA solution was essential to achieve the sensitivity reported in this communication. At least 1 log of sensitivity was lost when this step was omitted. The BSA apparently allows more efficient binding of the ASG. Also, the pH of the extracted food extracts proved to be very important. At acid pH, the level of nonspecific binding was increased, thus resulting in significantly less sensitivity. The best sensitivity was achieved at pH 8.0. %

E 3.0

2.0

10 0

10

1.0

100

STAPHYLOCOCCAL ENTEROTOXIN A (ng/mI)

FIG. 1. Standard curves illustrating detection by EIA of SEA in food extracts: mayonnaise, 0; milk, , hot dog, A. Each point is the mean of three replicates. The protocol is described in the text. To illustrate the absorbance values obtained from a sample containing only food juice (O ng of enterotoxin per ml), we arbitrarily placed a zero point in what is actually the 0.1 point on a log scale. P c 0.05 values (compared with blank) were: mayonnaise, 1.6 ng/ml; milk, 3.2 ng/ml; hot dog, 1.6 ng/ml.

TABLE 4. Summary of EIA of mayonnaise and hot dogs spiked with SEAa Sample Sample

Mayonnaise

Hot dog

found Recovery added SEA EIA NoSEA No. of SEAM) by SE(ng/g)

1 2 3

10 5 2.5

(ng/g) 9.4 4.6 2.1

o E

1 10 9.8 2 5 4.2 3 2.5 1.8 a Reaction times: 60 min for enterotoxin, for ASG conjugate, 30 min for substrate.

%

94 92 84 98 84 72 60 min

The vafue at which a sample becomes significantly positive in a binding test requires a degree of experimentation with known samples. When we apply a similar EIA procedure to detect antibodies in animal serum, we have found the sample-to-blank ratio a good indicator of positivity. For example, when looking for hog cholera antibodies, the absorbances of reaction product generated after a sample has been incubated with both hog cholera virus-infected and noninfected cells are compared; when the infected-tononinfected ratio is at least 1.5, we can say with 99% certainty that antibodies are present in that serum (10). In the present study, we have presented the sample-to-blank ratios in the data tables. However, we used a simple t test to distinguish the first sample in which there was 95% confidence of that sample reading being significantly higher than the blank. The sample-to-

ACKNOWLEDGMENTS We wish to thank Jack Leighty, U.S. Department of Agriculture, APHIS, MPIP, Washington, D.C., for both his support and useful discussions during the course of this work. This work was supported by U.S. Energy Research and Development Administration contract 8 W-7405-ENG.36, by the United States Department of Agriculture and the Energy Research and Development Administration under ERDA/USDA interagency agreement, and also by U.S. Public Health Service grant AI-11844 from the National Institute of Allergy and Infectious Diseases.

1.

LITERATURE CITED Bergdoll, M. S. 1970. Enterotoxins, p. 265-326. In T. C.

Montie, S. Kadis, and S. J. Ajl (ed.), Microbial toxins, vol. 3. Academic Press, Inc., New York. 2. Casman, E. P., and R. W. Bennett, 1965. Detection of staphylococcal enterotoxin in food. Appl. Microbiol. 13:181-189. 3. Collins, W. S., J. F. Metzger, and A. 0. Johnson. 1972. A rapid solid phase radio-immunoassay for staphylococcal B enterotosin. J. Immunol. 108:852-856. 4. Engvall, E., and P. Perlmann, 1972. Enzyme-linked immunosorbent assay, ELISA. mI. Quantitation of specific antibodies by enzyme-linked anti-immunoglobulin in antigen-coated tubes. J. Immunol. 109:129-135. 5. Genigeorgis, C., and J. K. Kuo. 1976. Recovery of staphylococcal enterotoxin from foods by affinity chromatography. Appl. Environ. Microbiol. 31:274-279. 6. Hebert, G. A., B. Pittman, R. M. McKinney, and W. B. Cherry. 1972. The preparation and physiochemical characterization of fluorescent antibody reagents. U. S. Department of Health, Education and Welfare, Public Health Service, Center for Disease Control, Atlanta, Ga. 7. Johnson, H. M., J. A. Bukovic, and P. E. Kauffmann. 1973. Staphylococcal enterotoxins A and B: solid-phase radioimmunoassay in food. Appl. Microbiol. 26:309-313. 8. Nakane, P. K., and A. Kawaoi. 1974. Peroxidase-labeled antibody, a new method of conjugation. J. Histochem. Cytochem. 22:1084-1091. 9. Reiser, R., D. Conaway, and M S. Bergdoll. 1974. Detection of staphylococcal enterotoxin in foods. Appl.

522

SAUNDERS AND BARTLETT

Microbiol. 27:83-85. 10. Saunders, G. C. 1977. Development and evaluation of an enzyme-labeled antibody (ELA) test for the rapid detection of hog cholera antibodies. Am. J. Vet. Res. 38:21-25. 11. Saunders, G. C., and E. H. Clinard. 1976. Rapid micromethod of screening for antibodies to disease agents using the indirect enzyme-labeled antibody test. J. Clin. Microbiol. 3:604-608. 12. Silverman, S. J., A. R. Knott, and M. Howard. 1968. Rapid, sensitive assay for staphylococcal enterotoxin and a comparison of serological methods. Appl. Microbiol. 16:1019-1023.

APPL. ENVIRON. MICROBIOL. 13. Van Weemen, B. K., and A. H. W. M. Schuurs. 1971. Immunoassay using antigen enzyme conjugates. FEBS Lett. 15:232-236. 14. Volier, A., D. E. Bidwell, and A. Bartlett. 1976. Enzyme immunoassays in diagnostic medicine. Theory and practice. Bull. W.H.O. 53:55-65. 15. Wolters, G., L. Kuijpers, J. Kacaki, and A. Schuurs. 1976. Solid-phase enzyme-immunoassay for the detection of hepatitis B surface antigen. J. Clin. Pathol. 29:873-879. 16. Zehren, V. L., and V. F. Zehren. 1968. Examination of large quantities of cheese for staphylococcal enterotoxin A. J. Dairy Sci. 51:635-644.