Proposed method for isolation of Escherichia coli 0157:H7 from environmental samples.

Introduction

Multiple cases of hemorrhagic colitis and hemolytic uremic syndrome (HUS) have been linked to the pathogen Escherichia coli, serogroup 0157:H7, present in meat and poultry products (1,2), raw milk, and community water supplies (3). Foods of bovine origin, especially ground beef, are the most frequently incriminated vehicles of infection. Additionally, there have been two waterborne outbreaks of infection induced by E. coli 0157:H7. The first, in Cabool, Missouri, was associated with the breakdown and associated repairs of an unchlorinated and aging sewerage and water distribution system (4). The second occurred in Portland, Oregon, during the summer of 1991, where a large number of visitors to a public swimming facility were stricken with E. coli 0157:H7, independent of food consumption (5). In a recent study performed in rural areas surrounding Philadelphia, Pennsylvania, abundant wildlife populations were linked to a reservoir contaminated with E. coli 0157:H7 (6). To date, the literature has concentrated on isolation of the organism from food or fecal matrices. Research focusing on the isolation of E. coil 0157:H7 in a water or soil environment, and the persistence in these milieu, is wanting. Coliform bacteria are commonly found in fide water, harbors, and bays, which are often badly polluted. Although E. coli and other lactose-fermenting bacteria do not constitute part of the normal intestinal flora of marine fish, if ingested, such facultative bacteria may survive for considerable periods of time (7). The potential for exposure to the E. coli 0157:H7 serogroup in a marine milieu (i.e. consumption of contaminated fish or shellfish, incidental ingestion of contaminated water through recreational usage of a harbor or bay) may be significant. The purpose of this study was to examine the utility of a procedure, adapted in part from Doyle and Schoeni (1), for detecting E. coli 0157:H7 in a spiked environmental sample. As the procedure is applied to other matrices, it is anticipated that this methodology may be useful for determining presence and survivability of E. coli 0157:H7 in unspiked receiving waters and soils.

Methods and Materials

Using a sterile glass jar and sampling spoons, approximately 1.5 kg of slurry was collected from a bay shore at high tide. Sample pH was 7.5, and salinity was calculated at 3.1% (7). The sample was secured just below the water surface, then divided into three 400 g groups. The first subsample ("high" group) was inoculated and mixed with 4.0 ml of an 18-hour pure culture of 0157:H7 in Trypticase soy broth (TSB)([10.sup.-2] dilution). Culture incubation temperature was 35 [degrees] C. The second subsample ("low" group) was prepared by adding 1 drop (50 microliters) of the same 18-hour culture to a 99 ml phosphate-buffered saline blank and shaken. A 0.1 ml aliquot was then transferred to another 99 ml dilution blank and shaken. A further 0.1 ml aliquot of this dilution was then added to the slurry and homogenized ([10.sup.-8] dilution). The third (unspiked) subsample was utilized as a negative control. Subsamples were cooled immediately to 4 [degrees] C and stored in the dark until tested for E. coli 0157:H7 on day 3 and day 20.

After 3 days and then 20 days of refrigeration, ten-fold serial dilutions of each sample were prepared to a [10.sup.-10] dilution. From each dilution, a 0.1 ml sample was pipetted (in duplicate) to MacConkey (BBL Microbiology Systems, Cockeysville, Md.) and MacConkey-sorbitol (Difco Laboratories, Detroit, Mich.) agar plates. Each aliquot was spread-plated until dry on the agar surface. Plates were inverted and incubated aerobically for 18-24 hours at 35 [degrees] C. The dilution range of interest produced between 50-100 well-isolated CFUs per plate. Colonies which were considered to be presumptive E. coli 0157:H7 were pink-red on MacConkey agar and clear on MacConkey-sorbitol media.

Modified vancomycin Trypticase soy broth (mvTSB) was prepared and dispensed in 96 well microtiter plates. The mvTSB medium was composed of (per liter) Trypticase soy broth (TSB, 30 g; BBL Microbiology Systems, Cockeysville, Md.), bile salts 3 (1.5 g; Difco Laboratories, Detroit, Mich.), and dipotassium phosphate, 12-hydrate (1.5 g; EM Science, Gibbstown, N.J.) (1). The medium was autoclaved and allowed to cool to room temperature before vancomycin (20 mg; Sigma Chemical Co., St. Louis, Mo.) was added. The pH of the mvTSB was 7.3. Colonies were picked from the agar plate to the microplate, incubated at 35 [degrees] C, then cross-phenotyped using a sterile 48-pin replicator (Ellard Instrumentation, Seattle, Wash.).

For every well of interest a 5 ml test tube of Trypticase soy broth was prepared. The tubes were inoculated with a sterile wooden applicator from each suspect well, then incubated at 35 [degrees] C for 24 hours. The tubes were then briefly agitated, a 0.2-0.3 ml volume of Kovac's spot indole reagent (Difco Laboratories, Detroit, Mich.) was added, and results were noted after 10 minutes. Because some strains of lysine negative Escherichia hermanii will cross react with 0157 sera; we employed the lysine decarboxylase test on candidate sorbitol nonfermenters and lactose fermenters (8): 13 x 100 mm serology tubes (VWR Scientific, San Francisco, Calif.) of lysine decarboxylase broth (test group) and decarboxylase base broth (control group) (Difco Laboratories, Detroit, Mich.) were prepared. Test and control tubes were inoculated using sterile wood applicators. Sterile mineral oil (Mallinckrodt Inc., Paris, Ky.) was added to make a 3-4 mm layer on the medium surface. Tubes were incubated for 24 hours at 35 [degrees] C, then positive (purple and turbid) and negative (yellow and turbid) tubes were noted. Lactose (+), sorbitol (-), indole (+), and lysine (+) colonies were serotyped.

The presence of the 0157 antigen was determined using a latex agglutination kit (Unipath Ltd., Ogdensburg, N.Y.), according to the manufacturer instructions. The presence of the H7 antigen was determined using motility media (Difco Laboratories, Detroit, Mich.), H7 antisera (Difco Laboratories, Detroit, Mich.), and a motility inhibition assay (2). Procedure note: Because of nonspecific autoagglutination, appropriate positive and negative controls are always included in the 0157 and H7 assays.

Results and Discussion

After 3 days and 20 days storage at 4 [degrees] C, the "high" sample, inoculated with a [10.sup.-2] cell density, produced 3 positives for E. coli 0157:H7 on both day 3 and day 20. The "low" sample, inoculated with a [10.sup.-8] density, yielded 2 confirmed positives on day 3 and 1 positive on day 20. No background levels of E. coli 0157:H7 were detected in the negative control. The aforementioned data was obtained from a total of 48 lactose (+), and 48 sorbitol (-) candidate E. coli 0157:H7 colonies from each original sample. Positive isolates were sent to an independent laboratory for confirmatory analysis and verified for the 0157 and H7 antigens. The method detected the contribution of inoculum rather than background levels of the organism. Though background 0157:H7 from the surrounding rural area may provide a seasonal contribution to the nearby bay, the (-) control did not confirm such a contribution during this investigation.

To date, the survivability of E. coli 0157:H7 in an environmental matrix has not been fully elucidated. From its recent history, some researchers have characterized this serotype as a relatively labile strain, as indicated by various "pre-enrichment" steps incorporated within existing methodologies (1,9). However, the results of this research indicate an organism that can survive for an unknown duration (at least 20 days) exposed to imperfect conditions without pre-enrichment of any kind. This observation agrees with recent literature on the subject (4,9). Glass, et al. (9) found that E. coli 0157:H7 is not significantly inhibited by NaCI below 8.5%, with cells showing vigorous growth at concentrations up to 2.5%. Moreover, the organism propagated in alkaline conditions up to pH 9.0. Ease of isolation is a function of background flora. Although relatively high levels of lactose (+) and sorbitol (-) organisms can obscure spiked samples, the indole and 0157 latex antisera tests effectively eliminated phenotypic false-positives from consideration.

Conclusion

These results suggest that waters and sediments exposed to intestinal wastes of warm-blooded animal origin may be an important reservoir and vehicle of E. coli 0157:H7 infections. In most waterborne outbreaks the causal agent is only a suspected pathogen, since it is typically not recovered from the suspected source (4). Culturing for 0157:H7 according to a standard methodology must become routine in clinical and biological laboratories.

In order to assess the range of applicability for the method, other environmental matrices should be examined. Future investigations will include the recovery of this pathogen from leachate and fresh water samples. Early indications are that this method may be effective in isolating the E. coli 0157:H7 serotype from unspiked environmental samples. These results lend credence to the hypothesis that waterborne outbreaks have been responsible in recent years for E. coli 0157:H7 infection. Moreover, environmental exposures are often overlooked by clinicians leading to inaccurate or unknown etiologies. This method provides environmental and public health professionals with a potential tool to further investigate the existence and survivability of this pathogen in environmental matrices influenced by fecal pollution.

References

1. Doyle, M.P., J.L. Schoeni (1987), "Isolation of Escherichia coli 0157:H7 from Retail Fresh Meats and Poultry," Appl. Environ. Microbiol. 53:2394-2396.

2. Riley, L.W., et al. (1983), "Hemorrhagic colitis associated with a rare Escherichia coli serotype," N. Engl. J. Med. 308:681-685.

3. Griffin, P.M., R.V. Tauxe (1991), "The Epidemiology of Infections Caused by Escherichia coli 0157:H7, Other Enterohemorrhagic E. coli, and the Associated Hemolytic Uremic Syndrome," Epidemiol. Rev. 13:60-98.

4. Swerdlow, D.L., et al. (1992), "A Waterborne Outbreak in Missouri of Escherichia coli 0157:H7 Associated with Bloody Diarrhea and Death," Ann Intern Med. 117:812-819.

5. Keene, W., J. McAnulty, K. Hedberg, G. Oxman, D. Fleming (1992), "A Swimming Associated Outbreak of Escherichia coli 0157:H7 Gastroenteritis," Abstract #95. 32nd Annual Meeting ICAAC.

6. McGowan, K.L., E. Wickersham, N.A. Strockbine (1989), Letter. Lancet ii:967.

7. ZoBell, C.E (1946), Marine Microbiology: A Monograph on Hydrobacteriology, p. 182-186. Chronica Botanica Co., Waltham, Mass.

8. U.S. Environmental Protection Agency (1978), Microbiological Methods for Monitoring the Environment, EPA-600/8-78-017.

9. Glass, K.A., J.M. Loeffelholz, J.P. Ford, M.P. Doyle (1992), "Fate of Escherichia coli 0157:H7 as Affected by pH or Sodium Chloride and in Fermented, Dry Sausage," Appl. Environ. Microbiol. 58:2513-2516.

Acknowledgements

The authors would like to thank the U.S. EPA for supporting development of this method under the Environmental Services Assistance Teams Zone 2 Contract #68D10135. Confirmational analysis was provided by the Washington State Public Health Laboratory, Seattle, Washington.

Dr. Phillip Tarr is the recipient of an American Gastroenterological Association/Blackwell Scientific PubliCations Research Scholar Award.

Disclaimer

Although the research described in this manuscript has been funded wholly or in part by the EPA Contract #68D10135 to ICF Incorporated., it has not been subjected to the Agency's review and therefore does not necessarily reflect the views of the Agency; no official endorsement should be inferred.

Dean W. Boening, M.S., Toxicologist, ICF Kaiser Engineers, Environment and Energy Group, Port Orchard, WA 98366.

Phillip I. Tarr, M.D., Div. of Gastroenterology, Children's Hospital Medical Ctr., and Dept. of Pediatrics and Microbiology, Univ. of Washington School of Medicine, Seattle, WA 98105.