Detection of Salmonella Strains and Escherichia coli O157:H7 in Feces of Small Ruminants and Their Isolation with Various Media

Salmonella strains and Escherichia coli O157:H7 were detectedin 17 and 5 small ruminants in Virginia, respectively, of 287tested. Background microflora interfered with the fecal analysis.The combination of Salmonella enzyme immunoassay (EIA) detectionand xylose-lysine-deoxycholate agar isolation was satisfactory.Modifying enrichment to a 1:100 dilution enabled effective E.coli O157:H7 detection by EIA and isolation by sorbitol-MacConkeyagar with cefixime-tellurite.

INTRODUCTION

Top

Introduction

Direct plating using selective media was found to be successfulin detecting and isolating Salmonella strains and Escherichiacoli O157:H7 from stools of diseased humans (8, 15, 18). However,an enrichment step was necessary for enhancing the detectionand isolation of target pathogens when healthy cattle with possiblylow levels of shedding were being evaluated (5, 6, 7, 9, 17,24). For rapid detection, enzyme immunoassays (EIA) and immunomagneticseparation systems were developed to screen bovine fecal samplesbefore pathogen isolation by plating enriched samples (4, 30).Comments on the methods for detection and isolation of pathogensin animal feces are available (4, 16, 19, 24, 27). Meyer-Brosetaet al. (16) stated that previous cattle studies may have missedfecal pathogens and misclassified herds as negative due to theuse of limited-sensitivity culture methods. Oberst et al. (19)found that current methods for recovering and identifying E.coli O157:H7 from cattle feces are inconsistent and are hinderedby their inability to specifically and rapidly detect smallnumbers of organisms from the complex and variable matrix. Moreover,Tutenel et al. (27) suggested performing more than one E. coliO157 test on fecal samples to avoid underestimating the incidenceof E. coli O157 in cattle.

Small ruminants, such as sheep and goats, are potential carriersof Salmonella and E. coli O157:H7 (1, 11, 14, 29). Althougha difference between the survivals of E. coli O157:H7 bacteriain ovine and bovine manure was noted (13), specific informationon the sensitivities of EIA and common plate assays for detectingand isolating these pathogens in feces of small ruminants islimited. The objectives of this study were to screen for thepresence of Salmonella and E. coli O157:H7 in the feces of smallruminants by validated EIA protocols for foods and to evaluatethe sensitivities of common selective agars for pathogen isolation(8, 15, 18, 21, 24).

(An abstract of this study was presented at the Ninth BiennialSymposium on Minorities, the Medically Underserved & Cancerin Washington, D.C., in March 2004.)

Sample collection.

Top

Sample collection.

For the microbiological survey, feces were collected from grazingsheep and goats. The animals (with ages ranging from young tomature) were predominately meat breeds raised under forage-basedsystems with limited concentrate supplementation. For the inoculationstudy, feces were collected from 20 healthy male sheep at theVirginia State University research farm. The 20 sheep were ofbreeding age and were provided with high-quality alfalfa haydaily. In all experiments, fecal samples were collected fromthe rectum with sterile gloves and refrigerated before testingwithin 24 h.

Sample inoculation.

Top

Sample inoculation.

Salmonella enterica serovar Typhimurium ATCC 14028, Salmonellaenterica serovar Montevideo ATCC 8387, and E. coli O157:H7 strainsATCC 700728 and ATCC 35150 were cultivated on tryptic soy agar(unless otherwise stated, all media were from BioPro, Bothell,Wash.) at 36°C for 24 h. Colonies were diluted in a phosphatebuffer to $~$ 10⁸ CFU/ml and confirmed by plate counts. Fecal samplesfrom the 20 male sheep were mixed, and 3 g of the mixture wasinoculated with 20 µl of diluted inocula to achieve estimatedinoculation levels of 10¹ through 10⁶ CFU/g. To evaluate theeffect of background microflora, 3 g of each fecal sample wasautoclaved and inoculated with E. coli O157:H7 to 10¹ CFU/g.All control samples tested negative for the target pathogens.

Microbial detection.

Top

Microbial detection.

For Salmonella detection, each sample (1 g) was preenrichedin 9 ml of buffered peptone water at 36°C for $~$ 20 h, followedby enrichment in Rappaport-Vassiliadis broth at 42°C for $~$ 18 h and postenrichment in mannose (M) broth at 36°C for $~$ 7 h before the Salmonella EIA (Tecra, Frenchs Forest, Australia)was performed. This enrichment protocol was established accordingto a recently approved AOAC method (method 998.09) for testingraw foods and high microbial load product with the EIA (26,31). The EIA-positive M broth samples (100 and 10 µl)were plated on xylose-lysine-deoxycholate (XLD) and xylose-lysine-tergitol4 (XLT4) media (Difco, Sparks, Md.) for isolation. A minimumof three black (or black-centered) colonies were inoculatedinto triple-sugar-iron and lysine-iron agar slants from eachplate and incubated for 24 h at 36°C. Isolates with positiveslant reactions were then tested for agglutination with SalmonellaO Poly A-I & Vi antiserum (Becton Dickinson and Co., Sparks,Md.). For E. coli O157 detection, each fecal sample (1 g) wasenriched in 9 ml of modified EC broth with 20 mg of novobiocin(Sigma, St. Louis, Mo.)/liter at 42°C for 24 h before theE. coli O157 EIA (Tecra) was performed. This enrichment protocolis recommended by the manufacturer for fecal samples and wasvalidated by AOAC for testing meat products (25). In the inoculationstudy, a 1:100 sample-to-enrichment-broth ratio (using 1-g samples)was also tested in order to detect samples containing E. coliO157:H7 at 10¹ CFU/g. The EIA-positive enrichment broth (100and 10 µl) was then plated on sorbitol-MacConkey agar(SMAC) and SMAC with 0.05 mg of cefixime/liter and 2.5 mg oftellurite/liter (CT-Supplement; Dynal Biotech, Lake Success,N.Y.) (SMAC-CT) for isolation. Up to 10 colorless colonies weretested for agglutination with E. coli O157 and E. coli H7 latextests (Remel, Lenexa, Kans.). Bacterial DNA extracted from EIA-positiveenrichment samples with a commercial kit (QIAGEN, Valencia,Calif.) was amplified by PCR with primers specific for O157(RfbE gene) (forward, CTACAGGTGAAGGTGGAATGG; reverse, ATTCCTCTCTTTCTCTGCGG)and H7 (FliC gene) (forward, TACCATCGCAAAGCAACTCC; reverse,GTCGGCAACGTTAGTGATACC) as previously reported (32). Amplifiedproducts were separated by agarose gel electrophoresis and visualizedby ethidium bromide staining. To determine background microbiallevels, appropriate dilutions (in 0.1% peptone) of each samplewere plated on standard method agar with 48 h of incubationat 36°C (for aerobic plate counts) or on acidified potatodextrose agar with 5 days of incubation at 25°C (for yeastand mold counts). E. coli counts were determined by a three-tube,most-probable-number evaluation according to the U.S. Food andDrug Administration method and statistical table (28). Afterincubation for 24 to 48 h at 36°C, a loopful of culturefrom a lauryl sulfate broth tube positive for gas productionwas transferred to EC-mug broth (EC broth with 4-methylumbelliferyl-ß-D-glucuronide).After incubation at 45.5°C for 24 to 48 h, tubes with gasproduction and fluorescence under long-wave UV light (336 nm)indicated positive E. coli results. The means and standard errorsof log₁₀ microbial counts were reported.

Microbiological survey results.

Top

Microbiological survey results.

Salmonella and E. coli O157 were detected by visual EIA withsamples from 17 and 7 small ruminants, respectively, of the287 tested (Table 1). Salmonellae were isolated exclusivelyin October, and one fecal sample tested positive for both pathogens.The presence of Salmonella and E. coli O157:H7 in the Virginiaherds was expected, as previous surveys conducted at differentlocations yielded similar results (1, 13, 20, 22, 23). Additionalstudies, with considerations for age, breed, and location, etc.,are needed to determine if the high prevalence of Salmonellafound in October is a significant trend in Virginia herds.

View this table:
[in this window]
[in a new window]
TABLE 1. Detection and isolation of Salmonella and E. coli O157 in fresh feces of sheep and goats from 12 Virginia farms

Salmonella organisms were isolated from 13 of 17 EIA-positiveenrichment broths by the use of XLD and XLT4 plates, with positiveconfirmation by biochemical and serological assays. Attemptsto isolate E. coli O157:H7 by using SMAC and SMAC-CT plateswere unsuccessful, except in one instance in which E. coli O157:H7was isolated by use of SMAC-CT with confirmation. For the EIA-positivesamples that were unsuccessful in isolation, PCR assays confirmedthe presence of the E. coli O157 gene in five of six samplesand E. coli O157 and E. coli H7 genes in four of six samples.Average background levels of microflora in the feces from 60sheep and goats included aerobic organisms at 7.2 ± 0.7log₁₀ CFU/g, yeasts and molds at 3.4 ± 0.7 log₁₀ CFU/g,and generic E. coli organisms at 5.5 ± 1.0 log₁₀ (mostprobable number)/g.

Results from inoculation study.

Top

Results from inoculation study.

Data in Table 2 indicate that visual EIA, when used with therecently approved AOAC method (method 998.09) involving Rappaport-Vassiliadisand M broths for sample enrichment (31), was effective for screeningfor the presence of all test levels (from 10¹ to 10⁶ CFU/g)of Salmonella serovar Typhimurium and Salmonella serovar Montevideoorganisms in sheep feces. Furthermore, the isolation sensitivityof XLD or XLT4 plates was improved (from $≥$ 10⁴ CFU/g) by about2 log₁₀ CFU/g when the enriched sample broth was used for EIA.These results show that the combination of visual EIA detectionand selective plate isolation (such as that with XLD and XLT4)with the enriched fecal samples is practical and effective forsmall-ruminant research.

View this table:
[in this window]
[in a new window]
TABLE 2. Detection and isolation of Salmonella and E. coli O157:H7 in artificially inoculated fresh feces of sheep

The presence of E. coli O157:H7 at low levels ( $≤$ 10² CFU/g) wasless likely to be detected by EIA. After enrichment, PCR identifiedthe presence of E. coli O157:H7 inoculated at the 10²-CFU/glevel consistently but at the 10-CFU/g level inconsistently.Isolating E. coli O157:H7 by streaking diluted feces on SMACand SMAC-CT plates was generally successful at inoculation levelsof $≥$ 10⁵ CFU/g and $≥$ 10³ CFU/g, respectively. Standard enrichmentdid not improve isolation with SMAC and SMAC-CT plates. Theresults from the survey also show inconsistent recoveries ofE. coli O157:H7 from the fecal samples, even when the EIA andthe PCR assay samples tested positive for the targeted antigensand genes. The isolation failures could be explained by thepresence of low levels of E. coli O157:H7 in feces that containhigh levels of background microflora, including sorbitol-positiveE. coli at about 5.5 log₁₀ CFU/g. The generic E. coli organismsalong with other unidentified colonies overcrowded the surfacesof SMAC or SMAC-CT plates, restricting the formation of identifiablecolonies from E. coli O157:H7 at lower levels. This overcrowdingis perhaps a problem for all types of selective agar plates,since no known E. coli O157-selective plates could effectivelyinhibit the growth of generic E. coli organisms (2, 3, 8, 15,18).

The results in Table 2 clearly show that E. coli O157:H7 maynot be effectively isolated from sheep feces by common plateassays unless its population is high (>10⁴ CFU/g). The EIAwith the standard fecal enrichment protocol is a more sensitivedetection method than the SMAC or SMAC-CT plate methods fordetecting the presence of E. coli O157:H7 in sheep feces. Dilutionof background microflora by modifying the sample enrichmentratio to 1:100 from the standard 1:10 enabled consistent EIAdetection and postenrichment SMAC-CT, but not SMAC, isolationfor three of three fecal samples inoculated with strains ofE. coli O157:H7 at 10¹ CFU/g (Table 3). Furthermore, eliminationof background microflora by autoclaving samples before the inoculationof E. coli O157:H7 enabled consistent EIA detection and postenrichmentisolation either by SMAC-CT or by SMAC. At a low level (10¹CFU/g) of inoculation, even PCR detection can be inaccurate(Table 2). Interference in detecting low levels of E. coli O157by PCR may be due to the relative growth of background fecalmicroflora or the presence of potential PCR inhibitors, suchas polysaccharides, bilirubin, and bile salts, commonly foundin animal feces (10, 12).

View this table:
[in this window]
[in a new window]
TABLE 3. Enhancement of E. coli O157:H7 detection and isolation sensitivities by reduction of background microflora in fecal samples from sheep

In conclusion, Salmonella and E. coli O157 were detected ina limited number of small ruminants in Virginia by the use ofconventional EIA enrichment and detection. The combination ofEIA and selective plate assays with conventional Salmonellaand improved E. coli O157 enrichment is relatively simple andinexpensive and may be easily merged (in medium preparationand sample incubation, etc.) with current food testing protocols(21, 26). These protocols are useful for on-farm evaluationwhen PCR or immunomagnetic separation systems are considerednot feasible or are not available.

FOOTNOTES

* Corresponding author. Mailing address: Virginia State University, Agricultural Research Station, P.O. Box 9061, Petersburg, VA 23806. Phone: (804) 524-6715. Fax: (804) 524-5186. E-mail: spao{at}vsu.edu.

Contribution of the Virginia State University Research Station(Journal Article Series no. 236).

REFERENCES

Top