Journal oflmmumdogical Method~. 152 (Itlt;2) 135-142 ~t~ 1992 Elsevier Science Publishers B.V. All rights reserved 11022-1759/t12/$115.1111

135

JIM 06362

Detection of Salmonella enteritidis in environmental samples by monoclonal antibody-based ELISA R.L. B r i g m o n a, S.G. Z a m h, G. B i t t o n ~' a n d S.R. F a r r a h i, Departments of " Em'irtmmental Engineering Sch'nces and i, Microbioh~gy and ('ell Science Unice~sity of bTorida, (;ainescilh.. FL 32611. USA

(Received 19 December 1991. revised received 21 February It)t)2. accepted 24 February I~)t)2)

We have developed a enzyme-linked immunosorbcnt assay (ELISA) using a monoclonal antibody (ASCII) for the detection of Salmonella enteritMis in environmental samples. ELISA was used to ~est for sensitivity and specificity of ASCII. 38 other species of bacteria, including 31 Salmonella species were included in cross-reactivity testing with ELISA. ASCII showed no reactivity with any other species tested. ASCll was found to be an lgGl specific for S. enteritMis lipopolysaccharide (LPS). The lower limits i'or S. enteritidis detection was 10 ~ cells/ml for pure cultures and in 10% sludge (w/v). Environmental samples (raw wastewater, wastewater effluents, mixed liquor and aerobically digested sludge) were obtained twice from five sites and ELISA tested for the presence of S. enteritidis. ELISA results compared to the American Public Health Association (APHA) method of Sabnonella detection were not significantly different (P > 0.05). The ELISA took 24 h for completion compared to 96-120 h for the A P H A procedure. Results demonstrate the reliability of the ELISA and, more importantly, provides a rapid means of detection of S. enteritidis in environmental samples. Key ~w~rds: Monochmal antibody; Sahnonella enteritidis; Environmental samples; ELISA

Introduction Salmonellae are bacterial pathogens of great concern from a public health viewpoint. Salmonella enteritidis, family Enterobacteriaceac, is a

Correspondence to: S.G. Zam, University of Florida. Department of Microbiology and Cell Science. 11153 McCarty Hall, Gainesville. FL 32610, USA. Tel.: (9041-392-1192; Fax: (904)-392-8479. i Florida Agricultural Experimentation Station, Journal Series Number R-II221 I. Abbretiations: MAB, monochmal antibody; pI3S, phosphate-buffered saline; ELISA, enzyme-linked immunosorbent assay; APHA, American Public Health Assc,iation; CDC. Center for Disease Control.

gram negative rod of the genus Salmonella which can cause infection in man and animals. While over 2(J0I) scrotypes of Salmonella have been established, a group of 10 serotypes accounted for 73% of the total human isolates identified in 1986 in the United States. S. enteritidis ranked second in this group of ten serotypes, accounting for 14% of the isolates (Center for Disease Control (CDC), 19871. Incidence of S. enteritidis as a cause of disease in humans has increased 3(X1% in 10 years (CDC, 1987). This increase is continuing and international in nature (Rodrigue et al., 199(I). It is estimated that there are from 400,000 to 4,0(10,000 Sabnonella infections annually in the United States (Cohen and Tauxe, 19861. Median duration of e~cretion time of Salmonella from those infected individuals is about 5 weeks

136

(Buchwald and Blaser, 1984). It can be estimated that there are at any one time approximately 401),000 people who are Sahnonella excretors in the USA. There are various methods for detection of Sabnonella which can take up to 96 hours for a complete test, depending on the method used, the initial concentration of Sahnonella, and physical characteristics of the sample (Harvey and Price, 1979). In recent years, there have been several methods designed to shorten the time for isolation or determination of the prese-.ce of Sahnonella in clinical, food, and environmental samples. Accurate detection of Sahnonella in environmental or mixed culture samples is particularly difficult and time-consuming due to sample variability. Those methods include: fluorescentantibody techniques (Thomason, 1971), radiometry (Stewart et al., 1980), DNA hybridization (Fitts ct al., 1983), enzyme-linked immunosorbent assay (Mattingly, 1984; Flowers et al., 1986), agglutination (Sveungsson and Lindberg, 1978), improved selective media (Perales, 1989), membrane filtration followed by culture and serological testing (Cerqueria et al,, 1986), radioimmunoassay (Brigmon, 1986), and immunomagnetic assays (Luk and Lindberg, 1991). The objective of these studies was the developmcnt of rapid, reliable, reproducible, and sensitive methods for the detection of salmonellae. However, some of the problems encountered are cross-reactivity, false positives and negatives, inconsistency, and cumbersome operation. S. enteritidis was the only salmonella serotype repeatedly isolated from sludge and sewage wastewater treatment plants i-. Fiorida (Farrah and Bitton, 1984). This report describes an ELISA procedure using a highly specific MAB to detect S. enteritidis in environmental samples. The ELISA was fast, sensitive, and reliable. The results were not significm~tly different from culture tcchniques and were obtained within 24 h.

Materials and methods

Bacterial strains All bacterial strains listed in Tablc 1, with the exception of Legionella pneumophilia, an environ-

TABLE I ELISA T E S T I N G O F ASCII F O R S P E C I F I C I T Y T O SELECT WHOLE BACTERIA

Bacteria

Absorbance (405 rim)

S. enteHtidis S. typhinutrium S. schottmulleri S. paratyphi S. choh'rasuis S. tennessee S. cerro S. litchfiehl S. derhy S. heidelberg S. t'td~ana S. gantinara S. newport

2.35 0.03 11.1)8 0.10 0. [ 5 11.114 0.07 0.13 I).14 0.08 0.02 0.[)4

S. n~hislaw S. Itlltellchell S. saintpaul S. harams

0.1]8 O.Ofi 0.09 0.04

0.04

S. djakarta

0.09

S. attatltm S. mississippi S. hartford S. ha~htr S. incerness S. sitlgapore S. harittingess S. urbana S. oranienburg S. I)mma S. git'e S. manhattan S. hraenderup Citrohacter fi'emufii Proteus mirabilis Vihrio parahaenudyticus Shigelhl dysenteriae Shigella boydi Legionella pnet#nophilia l'~sctwrichM coli Etltt'rohactt'r cloacae

0.08 0.03 0J)4 0.03 0.05 0.05 0.02 0.05 O,Ol 11.04 (1.10 (I.O0 0.113 0.(16 0.1)7 0.05 0.05 0.01 0.04

0.08 O.I)O

* Absorbance > 11.21)considered positivevalue. mental isolate, were obtained from the Department of Microbiology and Cell Science, University of Florida, Gainesville, FL, or from Dr. EIdert Hartwig, Florida State Public Health Laboratory, Jacksonville, FL. All media were commercial products of Difco Laboratories, Detroit, MI, except buffered yeast coated charcoal extract (BYCE), which was obtained from Remel, Lenexa, KS.

137 All stock cultures, with the exception of L. pneumophilia, were grown for 36-48 h in brain heart infusion (BHI) broth. Legionella was grown on BYCE slants for 36-48 h and then washed with 1% formalin in PBS. Cultures were centrifuged for 30 min at l(),()/)l}× g and the supernatants discarded. Bacteria were suspended in 1% formalin in 150 mM NaCI, 50 mM phosphatc buffer, pH 7.2 (PBS). After overnight refrigeration at 4°C, bacteria were washed three timcs with PBS. Samples of each washed culture wcrc reinoculated on media to insure total inactivation. Enumeration of bacteria was accomplishcd by comparison to a Mc.Carlanc standard and platc counts.

Immunization protocol B A L B / c mice were initially immunized in vivo by subcutaneous injection with 0.2 ml of Freund's complete adjuvant (Sigma Chetr:ical Co., St. Louis, MO) containing 2 × l0 s inactivated S. enteritidis (bioserotype enteritMis). This was repeated after 2 weeks. At week 4, the same mice were intraperitoneally injected with {).2 ml of Freund's incomplete adjuvant (Sigma Chemical Co., St. Louis, MO) containing 2 × l0 s inactivated S. enteritMis. At week 6, the mice were injected intravenously with 2 × l0 s inactivated S. enteritidis in PBS. At week 7, a blood sample was obtained from the tail vein and assayed for antibody production. The two mice having the highest antibody production were selected for hybridoma production. At week 12, these mice received a low dose ( < 103) of live S. enteritMis subcutaneously in 11.2 ml PBS. After I week, the splenocytes from the two mice were reimmunized in vitro with an adjuvant peptide, N-acetyI-L-alanyl-o-isoglutamine (Calbiochcm, La Jolla, CA) (Reading, 19861. Splenocytes were incubated in 20.(1 ml Dulbecco's modified Eagle medium (DMEM) containing 20% fetal bovine serum (FCS), supplemented with 5 × 10 -5 M 2-mercaptoethanol, 1 mM sodium pyrurate, 2 mM glutamine, 600 p,g of adjuvant peptide, 1 × 107 ultraviolet (UV)-inactivated S. enteritidis, 50 p,g/ml gentamicin, and 2.5 p g / m l amphotericin B. The suspension was placed in a 75 cm "~ tissue culture flask and incubated in a humidified 5% C O 2 / 9 5 % air incubator for 5 days.

Production of hybridomas Mouse mycloma cells, S P 2 / 0 - A G I 4 , were grown in DMEM supplemented with 10% FCS, 50 t~g/ml gcntamicin, 2.5 p g / m l amphotericin B, and 1.3 x 10 -~' M 8-azaguanine (Sigma Chemical Co., St. Louis, MO). After 2 days cultures were centrifuged at 7(}0 × g for 10 min and recovered cells were grown in DMEM with I(1% FCS. Colcemid (Dcmicoline, Sigma Chemical Co., St. Louis, MO), was added to the culture media (10 /zg/ml) 48 h prior to the hybridization procedure to enhance hybridoma production (Miyahara ct al., 1984). Myeloma cells and splenocytes from the hypcrimmunized mice were fused at a 1:10 ratio by a wlriation of the method of K6hler and Milstein ( 19751. 1 ml of 50% polyethylene glycol (PEG) (Sigma Chemical Co., St. Louis, MO) at 40°C was added over a I min interval to pelleted myeloma cells and splenocytcs and incubated for 1 min. DMEM was added to the suspension over a 3 min period at 10 m l / m i n . The cells were then centrifuged at 700 × g for 10 min and resuspended in 35 ml of D M E with 10% FCS fortified with hypoxanthine (1 × )0 ~' M), aminopterin ( 4 × ll) -'J M), and thymidine ( 1.6 × 10- 7 M) (HAT). The suspension was then distributed (100 tzl/wcll) into 4 - 9 6 well culture plates (NUNC, Denmark). The ceils were incubated as described previously and fed with 50 p.I of HAT every 5 days. After 2 weeks in HAT media, 100 tzl of supernatant was removed and 100 p l of DME supplemented with 10% FCS containing hypoxanthine (I × 10 ¢' M) and thymidine (l × 10 -7 M) (HT) was added to all wells Cctl cultures showing significant growth after 14-21 days were tested for antibody production to S. enteritidis by ELISA. Those cultures showing positive antibody activity were expanded and later retestcd. Cultures showing consistent results were cloned twice by limiting dilution, retested, and expanded to 30 ml cultures.

Purification of tile hybridoma antibody Hybridoma cells producing specific MAB were passagcd in B A L B / c mice. Briefly, pristane primed mice were injected with I(Y' hybridoma cells producing specific MAB and the ascitcs fluid collected. The antibody was precipitated

138 with 45% saturated ammonium sulfate and dialyzed against multiple chan~es of PBS over 48 h.

Enzyme-linked immunosorbent assay Formalinized washed bacteria were diluted to 10 ~ o r g a n i s m s / m l in carbonate-bicarbonate buffer (pH 9.6) and 100 p.I were added to 96 well immunoassay plates (Dynatech, Chantilly, VA). Positive controls containing only S. enteritidis and negative controls containing only P. mirabilis were included on each plate. Plates were incubated 1 h at 4°C. 50/~1 of PBS containing 0.5% glutaraldehyde were subsequently added to all wells for 15 min. The plates were then washed six times with PBS with 1% Tween 20 (Sigma, St. Louis, MO), and 200/.tl of bovine serum albumin (1%) (BSA)PBS were added to each well and incubated 1 h at room temperature. The plates were again washed and 100 /.tl of antibody (e.g., ascites preparation) was added to the wells. The plates were incubated at room temperature for 1 h. After washing to remove unbound antibody, I00 /.tl of a 1/1000 dilution in 1% BSA-PBS of affinity-purified alkaline phosphatase goat anti-mouse immunoglobulins (Organon Teknika, Malvern, PAl was added to each well. Plates were incubated for 1 h and then washed. 100 #1 of alkaline phosphatase substrate (1 m g / m l p-nitrophenyl phosphate, in diethanolamine buffer (Sigma Chemical Co., St. Louis, MOll, were added to each well. The plates were protected from light and read on a Titertek Multiskan (Flow Laboratories, McLean, VA) at 405 nra after 30 min. Antisera from immunized mice was used as a positive control and normal mouse sera was used as a negative control. A positive result was considered to be an OD of over 0.20 or three-fold above background. Five to ten-fold was usually the case.

Encironmental sampling Five sampling sites at the University of Florida sewage treatment plant, Gainesville, FL, were tested. Samples were final effluent, mixed liquor, raw sewage, aerobically digested sludge, and fresh sludge. Approximately 200 ml were taken at each site and processed in the laboratory within 45 min. Determination of presence of Salmonella by the American Public Health Association (APHA)

method and most probable number (MPN) was accomplished through the following steps. (1) Pre-enrichmem. Three tubes with 10 ml of nutrient broth were inoculated with 10 ml of fresh material, three tubes containing i0 ml of nutrient broth were inoculated with 1 ml of material, and a third set of three tubes with l0 ml of nutrient broth received 0.l ml of material. The samples were incubated 16 h at 39°C. (2) Enrichment. Pre-enrichment culture (l ml) was transferred to two sets of tubes, each containing 10 ml of selenite brilliant green broth for 4 h and 24 h, respectively. The 4 h set was used in the immunoassay procedure, and the 24 h set was used in step (3). (3) Inoculation. One Ioopful of the 24 h enrichment broth was streaked onto a plate containing one of the following media; brilliant green, salmonella-shigella, and hektoen enteric. At least two different media were used for each sample. The plates were incubated at 37°C for 18-24 h. (4) Purification. Portions from colonies resembling salmonellae species were streaked onto plates of Mac~2onkey agar for purification. (5) Characterization. Isolated colonies resembling salmonellae species were transferred to Kligler iron agar slants and incubated at 37°C for 24 h. Cultures giving reactions characteristic of salmonellae species were tested with polyvalent and group specific antisera (Difco Laboratories, Detroit, MI). Cultures that were agglutinated by salmonellae antisera were used to inoculate a series of biochemical tests in enterotubes (Roche Diagnostics, Nutley, N J).

hnmunoassay of em'ironmental samples After 4 h of incubation the contents of one set of enrichment tubes was pelleted by centrifugation at 10,000 rpm for 10 min and the supernatants discarded. Samples were diluted with ELISA buffer to the equivalent of a McFarlane ! standard. Duplicate 100 tzl samples were taken from each sample and tested by ELISA.

Sensiticity assays In order to test the lower threshold of the assay sensitivity, formalinized washed cultures of S. enteritidis were tested at dilutions between 103-10 ~ cells/ml in the ELISA.

Salmonella detection ht m:xed cuhures Due to the complex nature of environmental samples it would be unlikely to obtain a pure salmonellae culture at the enrichment stage of isolation. The ability of the assay to detect S. enteritidis added to mixtures of 10% sludge ( w / v ) which had tested negative for salmonellae was evaluated. Antigen determination Extraction of lipopolysaccharide (LPS) from S. enteritidis was accomplished by the hot phenol technique (Koga et al., 1985). Purified LPS was assayed in the EL1SA after diluting it to 5 p,g, 10 ttg, and 20 / z g / m l in carbonate-bicarbonate buffer. S. mhmesota and S. typhhnurium LPS (List Biological Laboratories, Campbell, CA)were used as negative controls.

_/.

"t 1.2]

R^2 = 0.946

1.0

m

/

m

0.8 0.6

~

0.4

~

02 O.

2

3

4

5

6

7

,

-

8

SalmoneDa enterlUdls log eel]s/ml in 10% sludge (w/v) Fig. 2. ELISA sensitivity testing of mixed culture with lot;• sludge and Sahnonella cnteritidi.~.

Results

Determination of specificity of monoclonal antibodies Hybridomas which tested positive in the ELISA were cloned, reidentified, cloned a second time, and reidentified again. The positive monoclonal antibodies were further tested against other

m

Sensiticit.v assays S. enteritMis can be detected at a minimum concentration of 1(15 cells/ml with the ELISA both in pure culture (Fig. !) and when S. enteritidis is mixed with 11)% sludge ( w / v ) (Fig. 2). Absorbance in the ELISA is highly correlated with S. entelqtidis cells/ml in both pure (R 2 = [).966) and mixed ( R - ' = 0.946) culture for the concentrations tested here. Logarithmic transformations were used to fit both curves.

^ = .

1

2

; ;

Sahnonella species and other bacteria found in the environment at a similar concentration. One MAB (ASCII) was used in this study because of its high specificity. ASCII was found to be of the IgGl type immunoglobin. Table I presents cr(~ssreactivity data of 1/10l) dilution of ASCII in PBS-BSA tested with other bacteria which could be found in environmental samples. No cross-reactivity resulted from testing with these bacteria including 30 other Sahnonella sp. Positive values are indicated by ODs over I).21).

~

~

~

~

Salmonella enteritldis logcells/ml Fig. 1. ELISA sensitivity testing of Sahnonella t'tllerilidis in pure culture.

Detection of Sahnonella in encironmental samples In the first sampling S. enteritidis was found at all sites sampled as indicated in Table il which gives an estimate of S. enteritMis number in original samples using the MPN technique (APHA,

1411

TABLE II

TABLE IV

DETECTION OF S. ENTERITIDIS IN WASTEWATER AND SLUDGE

ELISA TESTING OF LPS FROM S. ENTERITIDIS, S. MINNESOTA. AND S. TYPHIMURIUMAT CONCENTRATIONS OF 5, 10. AND 211 Atg/ml

Sample

R a w sewage Aerobic sludge

Mixed liquor Fresh sludge Final effluent

Most probable number (MPN) per 11111ml or g/dw "

Microorganism

LPS

Absorbance at 4115nm "

#g/ml

ASCII concentration 1:50

I : I(lO

1:500

5 10 20

1.83 1.89 1.92

1,80 1.88 1.84

11.65 1.53 1.58

5 10

0.01

0

I)

1)

0.02

0

0

11

20

0.111

I)

I)

5

0.01 0 11

I),Ol 0 0

0,01 0 0

Sampling date March 29. 1988

.June 2. 1988

APItA ~''" ELISAh"

A P H A ELISA

S. enteritidis

24,(11111 21(I 121I

II 11 II

I) (I 0

S. typhimurium

0 (I

I 1,0(ltl 210 121)

9 3

7 3

~'Grams as dry weight. h Test values were nt~l significantly dift;erent (P > I).051 when compared MPN values using Studcnl's t test (P = I).371). ¢ Test results from both tests wcrc highly correlated (P < 11.011 with respect to sampling site ( R: = 11.99).

S. mimlesota

1975). Test results fr,Jm both techniques were highly correlated ( R : = 11.99) with respect to sampling site as highest n u m b e r s were found in the raw sewage, and the lowest in the final effluent. Table Iii c o m p a r e s the two techniques with respect t,J positives and negatives. T h e r e was 82% a g r e e m e n t with the two tests. Results from the F L i S A technique indicated a higher n u m b e r of positive results than the A P H A m e t h o d (Table liD. The E L I S A test took approximately 24 h to complete, while the c u l t u r a l / s e r o l o g i c a l / b i o chemical m e t h o d ( A P H A ) took 96-120 h to complete. S. enteritidis was not found at any site on the second test using both techniques.

Determination The results ing of A S C I I LPS as shown

TABLE II1 DETECTION OF S. F,NTERITIDIS IN ENVIRONMENTAL SAMPLES BY CULTURE/SEROLOGICAL/BIOCItEMICAL (APItA) AND ELISA METHODS API4A

ELISA Positive " Negative

Positive h

Negative

23 I

5 16

" Samples were positive if absorbanee > 0.20 over controls. ~' Samples were positive if confirmed by cultural, serological and biochemical testing.

10 2O

" Absorbance > 0.20 considered positive.

o f antigen of the E L I S A confirmed the bindl g G l specifically to S. enteritidis in 'Fable IV.

Discussion The E L I S A technique indicated the presence of S. enteritidis in about 9.8% m o r e samples than the A P H A techniques. Based on the fact that these samples camc from polluted environments, the ELISA-positive e n r i c h m e n t s could have been were culture-negative due to ;.he over growth of salmonellae by o t h e r bacteria. T h e high correlation between absorbance and S. enteritidis (Figs. 1 and 2) makes the probability of false positives with E L I S A negligible. In addition E L I S A detects live, damaged, and dead S. o~teritidis. Most samples took 96 h to process using the A P H A technique but some needed 120 h as restreaking of plates was necessary due to overgrowth. This assay based on a highly specific MAB has proven to be accurate, reliable, and fast. While it did indicate more positive results than the traditional method, it has been established that standard microbiological t e c h n i q u e s often miss salmonellae in suspected samples (Harvey and

Price, 19791. The source of infection can bc difficult to determine by the low dose of Sabnonella required to initiate infection (Blascr and Newman, 1982; Buchwald and Blaser, 19841. Major Sahnonella disease outbreaks have occurred and the source later determined to be drinking water although the standard bacterial indicators, coliforms, were within the acceptable range (Greenberg and Ongerth, 19661. It has been suggested that Sahnonella monitoring would be a better indicator or coindicator of water quality than coliforms (Cherry et al., 19721. in Switzerland salmonellae were returned to the environment as a result of contamination of clarified sewage water and sludge which was used as fertilizer (Pagon et al., 19741. There has been a continuing increase of antibiotic resistant Sahnonellae strains as a cause of disease in the United States (Gill and Hook, 1966; Cohen and Tauxe, 1986). Salmonellae can multiply in both polluted and unpolluted river (Hendricks and Morrison, 1967) arid estuarine environments (Rhodes and Kator, 19881 at temperatures as low as 10°C, The presence and persistence of salmonellae in composted sewage sludge used to improve soil quality (Wray, 19751 containing antimicrobial resistant strains raises public health concerns. The original source of S. enteritidis infection can not always identified (St. Louis et al., 19881. The widespread occurrence of S. enteritidis in surface waters constitutes a public health problem the magnitude of which is obscured by inadequate diagnosis and under reporting of disease (Rodrigue et al., 199111. Hence, it is very important to d,wclop a simple method for detection of S. enteritMis in environmental samples. MABs have been used extensively to detect (Mattingly, 1984; Flowers ct al., 19861 salmonellae. The agglutinating abilities of MABs are useful reagents for the serological identification of salmonellae (Luk and Linberg, 1991). It has been recommended that an immunoassay utilizing MABs for detection of salmonellae in foods be used as a standard screening procedure (Flowers et al., 1986). In this study a murine IgGl producing hybridoma was developed producing MAB specifically directed against S. enteritidis LPS. This MAB

was tested for cross-reactivity against several bacteria as well as other Sabnonella sp. and has been tested using wastewater samples. Sludge mixed IllC/,~ ( w / v ) with S. enteritulis was used to assess the effect on the assay of different levels of competing organisms present in environmental samples. There was a decrease in absorbance in mixed culture (Fig. 2) compared to pure culture (Fig. 1) but detection was still evident. The sensitivity of the ELISA in this report is such that as few as I1)5 S. enteritMis cells/ml could be detected in both pure and mixed sampies. This scrotype was used h)r development of this assay duc to its repeated isolation from environmental samples ( F a r r a h and Bitton, It184) which was confirmed here as well as its increased incidence in the United States (CDC, 1987). If the incidence of S. enteritidis in man (Rodrigue, 19911)and food (CDC, 19871 continues t o increase it can be expected that more will be put into the environment through waste disposal. More research in this area of environmental testing in different systems utilizing the techniques described hcrc is in process. The method is both rapid and simple with a sensitivity which is sufficient for diagnostic and epidemiological investigations.

Acknowledgements We thank Mrs. Z.S. Zam, College of Medicine, University of Florida, Gainesville, FL, for providing mycloma cells and Dr. R.L. Tyndall, Oak Ridge National Laboratory, Oak Ridge, TN, for the L. pneumophilia, serotype I.

References American Public I lcalth Ass~leiation (1975) Standard Meth~ds for the Examination of Water and Wastewatcr, 14111 cdn. American Public tlealth Ass¢]ciation. New York. Blaser, M.J. and Newman. L.S. (1