Final Report: Detecting Fecal Contamination and Its Sources in Water and Watersheds

EPA Grant Number: R824782
Title: Detecting Fecal Contamination and Its Sources in Water and Watersheds
Investigators: Sobsey, Mark D.
Institution: University of North Carolina at Chapel Hill
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 1995 through October 31, 1998
Project Amount: $400,000
RFA: Water and Watersheds (1995) RFA Text |  Recipients Lists
Research Category: Watersheds , Water


The initial objectives of project were to further develop and evaluate and apply new and improved measures of human and animal sources of fecal contamination in water and watersheds. One approach was to be the direct measurement of microbial pathogens of documented public health risks using more reliable methods that detect potentially infectious organisms. The other approach was to be the measurement of candidate indicators of these microbial pathogens. Criteria used to judge the merits and reliability of the candidate indicators was to include their ability to predict the presence and concentrations of the pathogens relative to the sources of the organisms and their ability to predict the presence, persistence, transport and fate of the pathogens in response to natural environmental processes. In 1998, an additional objective was added to the project and the project period was extended to provide enough time to complete this objective. This supplemental project objective was to develop reliable, simple and economical methods to detect coliphages of fecal origin that reliably predict the risk of enteric virus contamination in ground water. Reliable, simple, rapid and economical methods to detect fecal coliphages in I 00-ml and I -liter volumes of ground water were to be adapted and refined from existing methods by determining and comparing the performance characteristics of several methods already available in the scientific literature. Acceptable methods used must have been: (1) simple enough to be done by any water microbiology laboratory; (2) rapid enough to provide results in as little as 24-48 hours; and (3) economical enough to cost <$100 per analysis.

Summary/Accomplishments (Outputs/Outcomes):

Methods were developed, refined and evaluated to detect the protozan pathogen Cryptosporidium, the bacterial pathogens Salmonella spp. and Yersinia enterocolytica, and indicator viruses (somatic and male specific coliphages) in water and wastes. As they were developed and refined, some of these methods were applied to the determination of pathogen and indicator levels in waters impacted by both point and non-point sources of fecal contamination in watersheds and aquifers. Samples were collected quarterly from fecal waste sources in watersheds and from a series of stream water stations in areas of two watersheds. One watershed area was impacted primarily by livestock agricultural wastes (hog and dairy cattle) and the other watershed area was impacted by multiple human and animal waste sources, depending on the creek or stream branch of the watershed and the location within the watershed. In addition, monthly samples were collected from an aquifer impacted by non-point sources of human and animal fecal contamination and examined for indicator viruses of fecal contamination using the new methods developed in this study. The extent to which the different candidate methods were able to detect fecal indicator viruses and the extent of fecal indicator virus contamination of the aquifer was determined.

Simultaneous Recovery and Detection of Viruses, Bacteria and Parasites in Water Using Hollow Fiber Ultrafilters. Reliable methods were developed for the simultaneous concentration of viruses, bacteria and protozoans from water using disposable, hollow fiber ultrafilters (see accompanying manuscript). Both tap water samples (50 liters) and raw, surface water samples (10 liters) were seeded with test organisms (Escherichia coli, coliphage MS2 and Cryptosporidium parvum) at high inputs and low inputs to determine the recovery efficiency of the system. The seeded water samples were recirculated through the filter using a peristaltic pump to concentrate to a final volume of <250ml. The sample was centrifuged to pellet the bacteria and oocysts. The supernatant was assayed for viruses and the pellet was resuspended in a small volume of phosphate buffered saline (PBS) for either direct enumeration of Cryptosporidium oocysts or further concentration prior to oocyst enumeration. MS-2 was assayed with the double agar layer (DAL) method using E. coli C3000 as the host. E. coli were enumerated from the pellet using membrane filtration and plating on MacConkey agar. The number of oocysts recovered was determined by microscopic counting after indirect immunofluorescent staining. For low level C parvum oocyst recoveries, a simple slide well method was developed for reliable microscopic enumeration of oocysts. This method gave oocyst recoveries of about 60 to 80 percent, which is much better than the membrane filter method previously specified by the U.S. Environmental Protection Agency (EPA) in the Information Collection Rule Method for parasite detection. The recovery efficiency of MS-2, E. coli, and C parvum averaged approximately 50 percent, 25 percent, and 50 percent, respectively. This procedure showed promise as a rapid, reliable, and economical means for the initial recovery of waterborne enteric pathogens, including Cryptosporidium oocysts.

Evaluation of USEPA Method 1622 for Recovery and Detection of Cryptosporidium in Turbid Stream Waters. During this study, the US EPA developed and reported a new and improved method to detect Cryptosporidium in water, known as Method 1622. This method consists of filtration, concentration, immunomagnetic separation, fluorescent antibody and DAPI counter staining, and microscopic evaluation. Because the method had not been evaluated for recovery and detection of C. parvum oocysts in stream water samples from watersheds, we evaluated the method for this purpose (see attached manuscript). Two recommended filters were compared for analysis of 11 stream water samples collected from various geographic locations throughout the United States. Oocyst recoveries from reagent water precision and recovery samples averaged 39 percent (S.D.? 13%) and 47 percent (S.D.? 19%) with the 293 mm diameter, polycarbonate membrane filter disk (Nuclepore) and the pleated, polyethersulfone membrane filter capsule, respectively. These recoveries from spiked reagent water samples are consistent with what had been reported in the literature by other laboratories. Replicate 10-L (2.6 gallons) stream water samples, unspiked and spiked with 100-250 oocysts, were tested to evaluate stream water matrix effects. Oocyst recoveries from the spiked stream water samples averaged only 22 percent (S.D.? 17%) and 12 percent (S.D.?6%) with membrane filter disk and the pleated capsule filter, respectively. C. parvum oocyst recovery efficiency was inversely correlated with stream water turbidity, showing that higher turbidity adversely impacted method performance. These results demonstrate that Cryptosporidium oocysts can be recovered from turbid stream waters using Method 1622, but recoveries are lower than from reagent-grade water. Concentrations of indicator bacteria (E. coli and Clostridium perfringens) also were evaluated in these stream water samples. Because few samples were oocyst-positive in the I O-liter volumes of unspiked water, relationships between detections of oocysts and concentrations of indicator organisms could not be conclusively determined.

Comparison and Ultrafilters and Capsule Filters for Recovery of Cryptosporidium in Stream Water by Method 1622. Because the currently recommended pleated capsule filter of US EPA Method 1622 gave relatively poor recoveries from stream waters, this capsule filter system was compared to the hollow fiber ultrafilter system we developed for primary concentration of C parvum oocysts in seeded reagent water and untreated surface waters (see attached manuscript). Water samples were otherwise processed according to the rest of the steps of Method 1622 (immunomagnetic separation, acid detachment of oocysts from immunobeads, application of oocysts to slide wells, immunofluorescent staining, counter staining and microscopic enumeration). Cryptosporidium parvum oocyst recoveries from seeded I O-L volumes of reagent water in precision and recovery experiments with filter pairs were 42 percent (S.D.=24%) and 46 percent (S.D.= 18%) for hollow fiber ultrafilters and pleated capsule filters, respectively. Mean oocyst recoveries in experiments testing both filters on seeded surface water samples were 42 percent (S.D.=27%) and 15 percent (S.D.= 12%) for hollow fiber ultrafilters and capsule filters, respectively. Although, C parvum oocysts; were recovered from seeded surface waters using the approved pleated capsule filter of Method 1622, recoveries were significantly lower and more variable than from reagent grade water. In contrast, the disposable, hollow fiber ultrafilter system was compatible with subsequent Method 1622 processing steps, and it recovered C parvum oocysts from seeded surface waters with significantly greater efficiency and reliability than the currently approved capsule filter.

Methods to Detect Salmonella in Water and Wastes. Alternative methods were developed and evaluated for the detection of Salmonella in water, sludge (biosolids) and liquid animal manure: traditional culture methods and molecular methods. Molecular methods involved membrane filtration, bacterial colony development, colony lifts and denaturation of nucleic acid, and nucleic acid hybridization with a synthetic, non-radioactive 16S rRNA oligonucleotide probe using chemiluminescent detection. For sludges, animal slurry and other samples with high suspended solids, the molecular methods involved pre-enrichment and enrichment in liquid medium, nucleic acid extraction, and dot hybridization using the same oligoprobe. For all samples, the traditional methods involved pre-enrichment and enrichment in liquid medium, plating on differential and selective agar media, biochemical testing and serology. A modified selective medium, lactose-combined tetrathionate (CTET), was used as an alternative to previously reported media for enrichment of Salmonella at 43 C and inhibition of other Enterobacteriaceae. For dot blotting, total RNA was extracted from 18-hour CTET broth cultures by TRI (RNA extraction) reagent and applied to nylon membranes. After hybridization for 2 hours, DNA-RNA hybrids were visualized by chemiluminescence. The oligoprobe reacted strongly with Salmonella spp., but it also weakly cross-hybridized with Citrobacter spp. CTET broth was used in an effort to inhibit Citrobacter growth and thereby prevent false positives. The molecular methods were initially successful in detecting and quantifying Salmonella in seeded samples of water and wastes. However, when these methods were applied to field water samples, they were unable to reliably detect and quantify Salmonella due to interference and false positive signals by Citrobacter and other non-Salmonella Enterobacteriaceae. Therefore, traditional culture methods had to be used to successfully detect and quantify Salmonella in environmental samples of water and wastes.

Recovery and Detection of Yersinia enterocolitica and Yersinia pseudotuberculosis in Environmental Water and Wastes by Cold-Enrichment. In order to determine the occurrence of pathogenic Y. enterocolitica (Y.e.) in fecally contaminated environmental waters, reliable culture methods are needed. Many culture media have been developed for the recovery of Y.e. from a variety of sources, but there has been no evaluation to establish their applicability to environmental water sampling. Thirteen broth media were evaluated for detection and recovery of Y.e. from spiked swine waste (swine are a major reservoir of Y.e.) by cold-enrichment at 4 C. Y.e. serogroups 0:3, 0;5,27, 0:8, 0:9 and Y. pseudotuberculosis were challenged with non-competitive (Yersinia only) and competitive (Yersinia plus antagonistic gram-negative rods) conditions. For non-selective broth media under noncompetitive conditions, BHIB and PEM possessed higher yields than PBS and TSB (see attached manuscript for media names). For selective broth media under non-competitive conditions, MTSB, PSB, and YSB possessed higher yields. Under competitive challenge, RVB possessed the highest yield and lowest antagonistic effect. Yersinia Selective Broth (YSB), a newly devised broth medium, effectively reduced competing gram-negative flora while retaining the ability to enrich Y.e. and Y. pseudotuberculosis to culture detectable levels (100 CFU/m1). No one cold-enrichment broth medium was best for all serogroups of Y.e or Y. pseudotuberculosis under these test conditions. Recovery from swine barn waste and water was optimized by the incorporation of a suite of nonselective and selective broth media.

Improved Detection of Fecal Contamination and Its Occurrence in Surface Waters of Watersheds (see attached manuscript). Surface water samples were collected quarterly for three years in Gaston County and Wake County, North Carolina, in watersheds having different levels of agricultural, residential, and urban development. Samples were collected at different stations in the Long Creek watershed (Gaston County) to investigate factors influencing the levels of enteric microbes potentially emanating from the following fecal waste sources: a daily farm, a municipal biosolids land application site and an urbanized region. Similarly, samples were collected in the Lake Wheeler watershed to investigate the potential impact of animal agricultural farm activities on the microbiological quality of nearby stream water. Geometric mean concentrations of various microbial indicators of fecal contamination (fecal coliforins, E. coli, enterococci, spores of Clostridium perfringens, somatic coliphages and male-specific coliphages) varied with land usage, sampling site, season, total suspended solids (TSS) measurements, and rainfall events. The strengths of these impacts on the observed microbial densities were variable, depending on the microbial indicator. Geometric mean concentrations of fecal coliforms, E. coli, and somatic coliphages were generally higher at sites impacted by animal and human inputs compared to rural residential and background stations. In addition, fecal coliform levels declined at the Gaston County dairy station after implementing best management practices for animal wastes. Runoff from a cattle pasture and feedlot at Lake Wheeler Road Farm contained high concentrations of microbial indicators, with enteric microbe densities approaching concentrations observed in raw animal wastewater. The levels of all microbial indicators increased significantly during rainfall events at all sites, as did TSS. Microbial indicators were positively correlated with TSS, but correlation strengths differed for base flow and storm samples. Compared to the other indicators, concentrations of enterococci and somatic coliphages were most influenced by season and were highest in the summer.

When fecal coliform results are compared to the NC standard for stream water quality, only 2 of 6 rural residence/background sites had geometric mean concentrations in compliance with this standard. The percentage of collected samples in compliance with this standard varied from 0 percent (livestock-impacted site) to 60 percent (rural/background site). The results of this study indicate that animal feeding operations (AFOs), biosolids land application areas and urban development can affect the microbial quality of nearby surface waters, though effects on baseflow water quality were not generally dramatic. However, stormwater data suggest that these waste sources can contribute substantial loadings of enteric microbes to surface waters during storm events.

Development and Evaluation of Methods to Detect Somatic and Male-specific Coliphages in Large Volumes of Ground Water and Other Waters. New and improved methods to detect somatic and male-specific coliphages of true fecal origin in large (0. 1 to I -liter) volumes of water were developed and evaluated (see attached manuscripts and report sections). The candidate methods for recovery and concentration/detection of coliphages were: (1) adsorption to and elution from 47 min diameter cellulose ester or electropositive fiberglass filters, followed by Double agar layer (DAL) plaque assay of the eluted viruses; (2) adsorption to 47 min diameter cellulose ester membrane filters and direct elution onto pre-poured host cell-agar media lawns for plaque assay; (3) overnight enrichment in broth media containing host bacteria, with detection either by a color change in the medium (from hydrolysis of X-Gal; "one-step enrichment") or by spotting from the overnight enrichment onto prepoured host cell-agar media lawns for development of lysis zones ("two-step enrichment"); (4) or a single agar layer (SAL) plaque assay (100-ml sample volumes only). The host bacteria evaluated for coliphage detection were E. coli strains C and CN 13 for somatic coliphages and E. coli Famp and Salmonella typhimurium WG49 for malespecific (F+) coliphages. Of the various recovery and concentration/detection methods tested, recoveries from samples of ground water seeded with natural coliphages populations of sewage origin were best using the cellulose ester membrane filter adsorption-elution-DAL plaque assay method (47% average recovery), the two-step enrichment method (100% average recovery) and the SAL method (86% average recovery). All of these methods were capable of reliably detecting as little as I or a few infectious coliphages in 0. 1 or I liter of ground water. Of the host bacteria tested, E. coli CN 13 was best for somatic coliphages and E. coli Famp was best for F+ coliphages. When the best of the candidate methods were compared by applying them to field samples of fecally contaminated surface water and ground water, the most sensitive method for coliphage detection was two-step enrichment (see attached manuscript).

Further studies were done to characterize the coliphages detected in fecally contaminated water. Of the four groups of male-specific RNA coliphages (Groups I, II, III, and IV), some Group II phages in field samples of water were not detected by the oligonucleotide probe we had previously developed. Therefore, an additional Group II probe was selected from nucleotide sequence data on Group II F+ RNA coliphages in GenBank. When the new Group II probe (1113) was used with the previous Group II probe (IIA), the combination detected all Group II F+ RNA coliphages (see attached manuscript). We found that it was possible to distinguish fecal from non-fecal coliphages on the basis of bacterial host range and growth temperature (see attached manuscript). Coliphages of fecal origin grew well at temperatures of 42-45 C and they had a narrow host range (most grow only on E. coli) (see attached manuscript). In field samples of water, the male-specific RNA coliphages appear to be the most specific in detecting fecal contamination and they grow at higher temperatures on only male-strains of E. coli or specially genetically constructed male-specific bacterial hosts.

Journal Articles on this Report : 1 Displayed | Download in RIS Format

Other project views: All 14 publications 1 publications in selected types All 1 journal articles
Type Citation Project Document Sources
Journal Article Simmons OD, Sobsey MD, Schaefer FW, Francy DS, Nally RA, Heaney CD. Evaluation of USEPA method 1622 for detection of Cryptosporidium oocysts in stream waters. Journal American Water Works Association 2001;93(1):78-87. R824782 (Final)
not available

Supplemental Keywords:

water, drinking water, watersheds, risk assessment, health effects, human health, mammalian, organism, pathogens, bacteria, discharge, ecosystem, indicators, aquatic, biology, ecology, epidemiology, monitoring, analytical, measurement methods., RFA, Health, Scientific Discipline, PHYSICAL ASPECTS, Water, Hydrology, Water & Watershed, Environmental Chemistry, Risk Assessments, Physical Processes, Biology, Drinking Water, Watersheds, monitoring, pathogens, contaminant transport, waterborne disease, agricultural waste, agricultural watershed, exposure, fecal contamination, other - risk assessment, microbial pathogens, municipal supply watershed, infectious organisms, municipal sewage effluent, microbial risk management, aquatic ecosystems, detecting fecal matter, salmonella, other - risk management, cryptosporidium

Progress and Final Reports:

Original Abstract
  • 1996
  • 1997
  • 1998 Progress Report